#  @cortensor Cortensor Cortensor posts on X about $cor, the new, core, status the most. They currently have [-----] followers and [---] posts still getting attention that total [------] engagements in the last [--] hours. ### Engagements: [------] [#](/creator/twitter::1802916496789929984/interactions)  - [--] Week [------] -42% - [--] Month [-------] +11% - [--] Months [-------] -25% - [--] Year [---------] +16% ### Mentions: [--] [#](/creator/twitter::1802916496789929984/posts_active)  - [--] Week [--] +9.30% - [--] Month [---] +1.70% - [--] Months [---] -23% - [--] Year [-----] +214% ### Followers: [-----] [#](/creator/twitter::1802916496789929984/followers)  - [--] Week [-----] -4.20% - [--] Month [-----] -5.60% - [--] Months [-----] +13% - [--] Year [-----] +18% ### CreatorRank: [-------] [#](/creator/twitter::1802916496789929984/influencer_rank)  ### Social Influence **Social category influence** [stocks](/list/stocks) 18.02% [finance](/list/finance) 13.51% [technology brands](/list/technology-brands) 9.01% [social networks](/list/social-networks) 7.21% [cryptocurrencies](/list/cryptocurrencies) 6.31% **Social topic influence** [$cor](/topic/$cor) #1, [the new](/topic/the-new) 15.32%, [core](/topic/core) #2970, [status](/topic/status) 13.51%, [the first](/topic/the-first) 10.81%, [testing](/topic/testing) 9.01%, [web3](/topic/web3) 7.21%, [module](/topic/module) #63, [in the](/topic/in-the) 6.31%, [inference](/topic/inference) 6.31% **Top accounts mentioned or mentioned by** [@el_cano498](/creator/undefined) [@cryptonormie50](/creator/undefined) [@01_choose](/creator/undefined) [@family_hedge](/creator/undefined) [@0xtwilty](/creator/undefined) [@cryptoninjapro](/creator/undefined) [@almamy1217047](/creator/undefined) [@djgriffith](/creator/undefined) [@bardieltech](/creator/undefined) [@bozdemir5269](/creator/undefined) [@mickey_tmnt](/creator/undefined) [@indespensionnl](/creator/undefined) [@01choose](/creator/undefined) [@itsmecryptality](/creator/undefined) [@zezuprof1](/creator/undefined) [@moltbook](/creator/undefined) [@openclaw](/creator/undefined) [@aixbtagent](/creator/undefined) [@widicrypto](/creator/undefined) [@corfaucetbota](/creator/undefined) **Top assets mentioned** [Cencora, Inc. (COR)](/topic/$cor) ### Top Social Posts Top posts by engagements in the last [--] hours "🛠 DevLog Dynamic Pricing Ready Across All Testnets Dynamic pricing is now wired up end-to-end on Testnet-0 Testnet-1 and Testnet-1a. 🔹 SessionPaymentTable per-network Global Pricing Table and Session Pricing Breakdown UIs are all deployed on Testnet-1 / Testnet-1a matching Testnet-0. 🔹 Well run manual user task e2e tests on Testnet-1 and Testnet-1a to confirm everything behaves as expected under real sessions. 🔹 Dynamic pricing is not yet "turned on" globally - we can flip a single switch and adjust the pricing tables during Testnet Phase #2 and Phase #3 to tune rates weights and behavior" [X Link](https://x.com/cortensor/status/2009925321462821365) 2026-01-10T09:48Z [----] followers, [----] engagements "🛠 DevLog Testnet-1a Router Live + First Automated Jobs Weve wired up the dedicated router for the self-managed $COR #L3 (Testnet-1a) and pointed the first automated jobs at it. 🔹 New Router Endpoint (Testnet-1a / Self-Managed #L3) - REST entrypoint: - Currently running [--] automated user-task jobs against Testnet-1a. - Well increase job count and patterns as more nodes join the Testnet-1a node pool. 🔹 How Well Use Testnet-1a This Phase - Focus automated load on self-managed #L3 while keeping Testnet-0 / Testnet-1 as known-stable baselines. - Use the runs to explore #L3 levers: - fee / gas" [X Link](https://x.com/cortensor/status/2013211710304534630) 2026-01-19T11:27Z [----] followers, [---] engagements "🛠 DevLog 30-Day Window to Decide #L3 Direction With Testnet-1a (self-managed $COR #L3) now under automated load were giving ourselves a [--] day window to make the #L3 call: stay with RaaS or fully lean into self-managed. 🔹 What Were Evaluating - If self-managed L3 continues to run smoothly under load (network tasks + user tasks + fee paths) - well decommission the RaaS L3 and move forward with self-managed COR L3 as the primary path. 🔹 Why This Is Plausible Now - #Arbitrum #Orbit tooling has matured a lot compared to when we first experimented [--] years ago. - Operational + financial profile" [X Link](https://x.com/cortensor/status/2013227018255736919) 2026-01-19T12:28Z [----] followers, [---] engagements "🛠 DevLog Self-Managed L3: Connectivity Findings & Next Steps Recent stress tests on Testnet-1a (self-managed $COR #L3) surfaced some very practical crossdata center connectivity limits in our setup. 🔹 What Were Seeing - DA (Data Availability) & Sequencer are running in one data center - RPC is running in a different data center - Under sustained load network hiccups between those DCs can: - Disrupt #L3 #L2 data sync - Occasionally force DA resyncs when connectivity between RPC and DA/Sequencer becomes unstable 🔹 What Were Planning - Co-locate core L3 components - Move RPC into the same DC" [X Link](https://x.com/cortensor/status/2013843004780593419) 2026-01-21T05:16Z [----] followers, [---] engagements "Staking APR Adjustment Reminder (Feb 28) Quick reminder that the next scheduled APR step-down for $COR staking goes live on Feb [--] [----] as part of the long-term sustainability plan. 🔹 New APR (from 2/28/2026) - Pool 3: 10% APR - Pool 2: 5% APR - Pool 1: 2% APR Lock periods + all other staking conditions stay the same. This follows the gradual schedule as the network matures and we move into the next phases of Testnet Mainnet. More details: #Cortensor #Staking #COR #DePIN https://docs.cortensor.network/community-and-ecosystem/staking-pool-overview#apr-breakdown 📆 Staking APR Adjustment Next" [X Link](https://x.com/cortensor/status/2014679532234834270) 2026-01-23T12:40Z [----] followers, [---] engagements "🛠 DevLog Testnet-1a & Testnet-1b Status Quick #L3 self-managed check-in. 🔹 Testnet-1a Load & Stability - [--] automated jobs are now running on Testnet-1a (self-managed $COR #L3). - So far things look stable after the move to dedicated servers. - As this holds through the week well ask more node operators to migrate nodes over and gradually increase load. - Live stats: 🔹 Testnet-1b Migration Ongoing - Testnet-1b (internal experimental #L3) is in the middle of its migration to the new dedicated setup. - This environment will stay our breakable lane for DA/Sequencer config experiments and ops" [X Link](https://x.com/cortensor/status/2014888039030653060) 2026-01-24T02:28Z [----] followers, [----] engagements "🛠 DevLog L3 Self-Managed Data Growth (Sequencer Size) Over the last 2448h on the self-managed $COR #L3 weve gotten a clearer picture of how fast sequencer data grows under real traffic. 🔹 What Were Seeing - Current self-managed L3 (Testnet-1a / 1b) sequencer data is already around 20GB after sustained blocks + TXs. - It confirms that a long-running L3 will accumulate significant state as we push more volume and stress tests. 🔹 What Well Explore Next - Track data growth vs TX count / block rate / DA setup more systematically across Testnet-1a and 1b. - Experiment with pruning snapshots and" [X Link](https://x.com/cortensor/status/2015386303584747717) 2026-01-25T11:28Z [----] followers, [---] engagements "🔹 DevLog Data Mobility & Data Module Import/Export (Early Design Thoughts) As we see real data growth on the self-managed #L3 its becoming clearer that storage growth itself isnt the real problem - data mobility is. 🔹 What Were Thinking About - Instead of worrying only about raw disk size were focusing on how to move and reshape state cleanly over time. - The idea is to support import/export for core data modules so we can migrate or hardfork infra without losing the logical network state. Target data modules for import/export design: - Session Data user sessions configs runtime state -" [X Link](https://x.com/cortensor/status/2015563260041613761) 2026-01-25T23:11Z [----] followers, [----] engagements "🛠 DevLog #L3 Data Growth vs Data Mobility (Testnet-1a) Over the last [----] hours on Testnet-1a (self-managed #L3) we finally have some rough numbers on data growth: 🔹 What Were Seeing - Sequencer side is growing around [--] GB/day. - Thats roughly [---] GB/month about [---] TB/year at the current pace. - Most of this is L2 sync / rollup data not pure L3 chain state we fully control. 🔹 Why Storage Isnt the Real Bottleneck - With modern SSD pricing even setups like [--] TB SSD [--] (RAID / multi-disk) per server are very doable. - We can also lean on high-throughput instances (e.g. AWS-type configs)" [X Link](https://x.com/cortensor/status/2015675123089780877) 2026-01-26T06:36Z [----] followers, [---] engagements "🗓 Weekly Focus Phase #2 Stability #L3 Load Expansion & Ranking Refinements This week is focused on keeping Testnet Phase #2 stable while we push heavier load onto the self-managed #L3 track refine validator ranking surfaces for future SLA filters and resume agent-facing endpoint work. 🔹 Testnet Phase #2 Stability & Support - Continue active monitoring and support across all environments (routing miners validators oracles indexers dashboards). 🔹 Testnet-1a Load Expansion + Storage Monitoring - Move more nodes from Testnet-0 Testnet-1a to increase L3 capacity. - Run more manual + automated" [X Link](https://x.com/cortensor/status/2015688030212997547) 2026-01-26T07:27Z [----] followers, [----] engagements "Monthly Node Rewards January [----] (Initial Batch Sent) We've just sent out the first round of $COR rewards to [--] active nodes as part of our monthly node reward program. 🔹 Rewards will be distributed in batches 🔹 More batches rolling out over the next few days 🔹 Recognizing uptime stability & contribution 💬 Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards December [----] (last Batch Sent) Another [--] active nodes" [X Link](https://x.com/cortensor/status/2015710197642351086) 2026-01-26T08:55Z [----] followers, [---] engagements "Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [--] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Initial Batch Sent) We've just sent out the first" [X Link](https://x.com/cortensor/status/2015912214516138477) 2026-01-26T22:18Z [----] followers, [---] engagements "🛠 Devlog Testnet-0 Snapshot & Shifting Load to Self-Managed #L3 Well be taking the next Testnet-0 snapshot for monthly node rewards shortly. Once thats done well start asking node ops to migrate more nodes over to Testnet-1a (self-managed #L3) so we can focus stress tests there. 🔹 Current Testnet-1a State - Explorer: - Stats: 1.8M tx processed 889k blocks 0.5s average block time Daily transactions now in the 350k range 🔹 Next 24h & Node Split Plan - If Testnet-1a stays stable over the next [--] hours well shift from a 50/50 node split (Testnet-0 vs Testnet-1a) to roughly 30% / 70%. - Goal:" [X Link](https://x.com/cortensor/status/2015933726480597261) 2026-01-26T23:43Z [----] followers, [---] engagements "📣 Node Ops Help Us Load-Test Self-Managed #L3 Were shifting more traffic to Testnet-1a (self-managed L3) and would like node operators to rebalance their nodes so that roughly: - 30% of your nodes stay on Testnet-0 (#Arbitrum #Sepolia) - 70% of your nodes run on Testnet-1a (self-managed $COR #L3) This lets us keep Testnet-0 as a stable baseline while really stress testing the self-managed L3 stack (sequencer DA RPC infra) on Testnet-1a. Dashboards: - Testnet-0: - Testnet-1a: #Cortensor #NodeOps #TestnetPhase2 #L3 #DePIN #AIInfra https://dashboard-testnet1a.cortensor.network" [X Link](https://x.com/cortensor/status/2015951928640077830) 2026-01-27T00:56Z [----] followers, [---] engagements "🛠 DevLog /delegate & /validate Iteration with Updated Corgent Endpoints Quick update as we keep iterating on /delegate and /validate on top of Corgent now with the correct router endpoints for both Testnet-1a and Testnet-0 and a clear path into the #ERC8004 ecosystem. 🔹 Testnet-1a Self-Managed L3 (Primary Playground) Internal (Cortensor): - MCP: - REST: External (Bardiel): - MCP: - REST: 🔹 Testnet-0 Baseline Environment - MCP (Cortensor): - MCP (Bardiel): 🔹 What Were Focusing On - Driving more real /delegate flows through Testnet-1a to refine routing retries and tool usage on self-managed" [X Link](https://x.com/cortensor/status/2016069068525339120) 2026-01-27T08:41Z [----] followers, [---] engagements "🛠 DevLog GLM-4.7-Flash Now Live Weve finished wiring in GLM-4.7-Flash and its now ready for wider use across the network. PR: 🔹 What Were Doing - GLM-4.7-Flash support is fully implemented images are built and tested and the PR is ready to use. - A few node operators have already tried it and early feedback looks solid. - Well now start rolling it out more broadly as an additional high-performance model option. 🔹 Why It Matters - GLM-4.7-Flash is extremely strong for agentic coding + tool-calling tasks - in many cases noticeably better than typical OSS baselines. - This gives Cortensor a" [X Link](https://x.com/cortensor/status/2016103332704846233) 2026-01-27T10:57Z [----] followers, [---] engagements "Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [--] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just" [X Link](https://x.com/cortensor/status/2016247543051780151) 2026-01-27T20:30Z [----] followers, [---] engagements "🛠 DevLog Corgent #ERC8004 Agents Wired for Testnet-0 & Testnet-1A Weve finished wiring Corgents external #ERC8004 agents so we can iterate on /delegate and /validate flows from the broader #ERC8004 ecosystem. 🔹 Testnet-0 Corgent Trust Oracle - Sepolia agent: - Base Sepolia agent: 🔹 Testnet-1A Corgent Trust Oracle (Testnet1A) - Sepolia agent: - Base Sepolia agent: These agents are backed by the Corgent #MCP endpoints on Testnet-0 / Testnet-1A and will be used for external /delegate + /validate experiments as we refine Corgent as a shared trust oracle for other agents and apps. #Cortensor" [X Link](https://x.com/cortensor/status/2016349538098020500) 2026-01-28T03:16Z [----] followers, [----] engagements "Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [--] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just" [X Link](https://x.com/cortensor/status/2016429611593462147) 2026-01-28T08:34Z [----] followers, [---] engagements "🛠 DevLog Corgent Primary #ERC8004 Agents + Bardiel #Virtual Layer We now have two layers of #ERC8004 agents wired into Cortensor with Corgent as the primary trust oracle and @BardielTech as the more #Virtual / #ERC8004facing layer. 🔹 Corgent Primary Trust Oracle Agents Testnet-0 Corgent (Trust Oracle) - Base Sepolia: - Sepolia: Testnet-1A Corgent (Trust Oracle Testnet1A) - Base Sepolia: - Sepolia: These are backed by the new router #MCP and #x402 endpoints on Testnet-0 / Testnet-1A and will be the main surface for /delegate + /validate flows as we refine Corgent as a shared trust oracle for" [X Link](https://x.com/cortensor/status/2016457286659793403) 2026-01-28T10:24Z [----] followers, [----] engagements "🛠 DevLog Refined Quantitative / Qualitative Rank Tables (Now on Testnet-0 & Testnet-1A) Weve rolled out the updated validator ranking views to both Testnet-0 and Testnet-1A. 🔹 Whats New - Refined Quantitative Rank Metrics (PoI) and Qualitative Rank Metrics (PoUW) tables with clearer scores + status labels. - Better handling of inactive / never-validated nodes so they dont pollute the healthy set. - Dialog header now shows quick stats (validated / online / inactive active score range) so operators can read network health at a glance. 🔹 Where To See It - Testnet-0 Quantitative: - Testnet-0" [X Link](https://x.com/cortensor/status/2016482583790665785) 2026-01-28T12:04Z [----] followers, [---] engagements "Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [---] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just" [X Link](https://x.com/cortensor/status/2016633936227815668) 2026-01-28T22:06Z [----] followers, [---] engagements "🛠 DevLog Self-Managed #L3 Infra-as-Code (Early Assessment) 🔹 What Weve Proven So Far - Testnet-1A (self-managed L3) is now past 2M tx and 1M blocks running stably under load. - Most of this was brought up and managed via manual setup and config over the last [--] weeks. 🔹 What We Started Today - Walked through the full L3 stack (Nitro DA nodes Blockscout stats proxies monitoring) to: - identify all manual steps - map required configs + secrets and - understand dependencies between components (L2 RPC DA cluster sequencer explorer etc.). 🔹 Where Were Headed - Explore turning self-managed L3" [X Link](https://x.com/cortensor/status/2016795931287634045) 2026-01-29T08:49Z [----] followers, [----] engagements "🛠 DevLog First End-to-End User Task Fee Claims on Testnet-1A 🔹 What We Tested - Started exercising the last mile of the payment flow: claiming user task execution fees from completed sessions. - Ran this on Testnet-1A (self-managed L3) where weve been pushing sustained load tests so node ops actually have real execution fees to claim. 🔹 Early Result - First end-to-end claim of user task execution fees appears to be working as expected on Testnet-1A. - This confirms that the path from: - user task execution validator confirmation node credit claimis now wired through the refactored" [X Link](https://x.com/cortensor/status/2016845271469031852) 2026-01-29T12:05Z [----] followers, [----] engagements "🛠 DevLog Corgent & Bardiel Registered as #ERC8004 Mainnet Agents 🔹 Whats New - #ERC8004 mainnet went live today and weve registered both #Corgent and @BardielTech as mainnet agents so theyre visible in the emerging ecosystem. - For now both are still backed by our Testnet-1A router stack under the hood while we continue to iterate on /delegate and /validate flows during the current testnet phase. 🔹 Mainnet Agent Entries - #Corgent (Primary Trust / Validation Agent) - Ethereum mainnet agent: - #Bardiel (Virtual-Native / Validation & Delegation Surface) - Ethereum mainnet agent: 🔹 Whats" [X Link](https://x.com/cortensor/status/2016982164529914206) 2026-01-29T21:09Z [----] followers, [----] engagements "🛠 DevLog Testnet-1A #L3 Focus Shift (Most Nodes Moving Over) 🔹 Network Status - Weve asked node operators to move all nodes to Testnet-1A keeping only one node on Testnet-0 (L2) as a comparison / fallback. - Testnet-1A is now our primary environment for self-managed L3 benchmarking and monitoring. - Current L3 stats: 1.3M blocks 2.7M tx and 230K daily tx and growing. 🔹 What Were Doing on Testnet-1A - Benchmarking L3 behavior under higher more realistic load (routing validation NodePool utilization). - Tuning reward logic so better-performing nodes are more clearly recognized /" [X Link](https://x.com/cortensor/status/2017193335648391563) 2026-01-30T11:09Z [----] followers, [---] engagements "🛠 DevLog Node Performance / Reward Config Refinement (Now Live on Testnet-1A) 🔹 Weve pushed the updated Node Performance & Reward view to the Testnet-1A dashboard and started using it as the primary way to shape monthly node incentives. 🔹 Live page (Testnet-1A): 🔹 New config knobs - Model Count Threshold minimum number of supported models before a node is considered for rewards. - Min Level (base reward) only nodes at or above a target level earn the base payout. - Must Model Indexes required model indexes that must be online to qualify (already wired into enforcement). - Level Bonus" [X Link](https://x.com/cortensor/status/2017432157501984873) 2026-01-31T02:58Z [----] followers, [---] engagements "🧪 Testnet Phase #2 - 2-Week Check-In (Jan [--] 2026) Its been [--] weeks since Phase #2 kicked off (Jan 17). Theme so far: stable baseline pushing more traffic to self-managed L3 and tightening the trust + routing + incentives loop. ✅ What Happened (So Far) 🔹 Self-Managed #L3 Is Now the Main Test Bed - Testnet-1A (self-managed L3) is at 1M blocks / 2M tx averaging 300K tx/day now running on dedicated servers with DA sequencer and RPC co-located. - Most nodes are being moved from Testnet-0 Testnet-1A targeting roughly 30% (T0) / 70% (T1A) for Phase #2 load. 🔹 L3 Ops Fees & Data Growth Mapped Out" [X Link](https://x.com/cortensor/status/2017551617114182112) 2026-01-31T10:52Z [----] followers, [---] engagements "🗓 Weekly Recap Phase #2 Stability Testnet-1a Load Shift & Ops Improvements This week we completed most planned Phase #2 work with the majority of activity and load now shifted to Testnet-1a (self-managed L3). 🔹 Testnet Phase #2 Stability & Support - Continued active monitoring across routing miners validators oracles indexers and dashboards. - Followed up on abnormal node states with NodeOps; overall Phase #2 stability improved. 🔹 Load Shift to Testnet-1a (Self-Managed L3) - The majority of nodes and testing activity moved from Testnet-0 to Testnet-1a. - Ran additional manual + automated" [X Link](https://x.com/cortensor/status/2017728451197997522) 2026-01-31T22:35Z [----] followers, [----] engagements "🛠 DevLog Testnet Node Capacity Market Conditions & Next L3 Steps 🔹 What Were Doing - Based on community feedback + current market conditions were keeping testnet node capacity intentionally moderate and adjusting it gradually instead of pushing for max scale right now. - Upscaling isnt the bottleneck - spinning up more capacity is straightforward once the timing + market are right. 🔹 What Weve Learned So Far - Last phases (plus current Phase #2) showed the stack can scale: more nodes + more traffic mostly surfaced minor bugs and SLA tuning gaps not hard infra limits. - That means the best" [X Link](https://x.com/cortensor/status/2017807591598850219) 2026-02-01T03:49Z [----] followers, [---] engagements "The quiet lesson from #OpenClaw / #Moltbook-style agent stacks: inference is getting cheap but running agents 24/7 is not. As we spin up fleets of always-on agents Jevons kicks in: lower per-call cost more calls more tools more orchestration and suddenly the real bottlenecks are: - Infra bills (API + hosting + storage) - Trust in what agents did while you werent watching - Long-term memory + state you dont want to recompute every time Cheap inference doesnt end demand it moves value from raw tokens to: ✅ verification / proofs ✅ scheduling & routing ✅ durable memory + context sharing. Thats" [X Link](https://x.com/cortensor/status/2017877615869354472) 2026-02-01T08:28Z [----] followers, [----] engagements "🛠 DevLog CorClaw Agent Economy Experiment (#OpenClaw + #Moltbook + #Clawnch) 🔹 What Were Actually Testing - OpenClaw is the local agent stack were experimenting with. - Moltbook is the agent social network where CorClaw / Cortensor agents can post follow and be discovered: - CorClaw profile: - Cortensor community: - Clawnch is a separate agent-powered launchpad that CorClaw used to deploy a small test token on Base as an on-chain fee stream: - Clanker view: 🔹 Why This Matters - Always-on agents are fun in theory but in practice 24/7 inference is still expensive especially at scale. -" [X Link](https://x.com/cortensor/status/2018106163708170345) 2026-02-01T23:36Z [----] followers, [----] engagements "Everyone loves the agent economy idea. Autonomous agents managing your work Bots trading on your behalf AI companions running 24/7 in the cloud On paper it looks inevitable. In practice it runs http://x.com/i/article/2018140779068395520 http://x.com/i/article/2018140779068395520" [X Link](https://x.com/cortensor/status/2018150602614907078) 2026-02-02T02:32Z [----] followers, 37K engagements "🗓 Weekly Focus Phase #2 Hardening SLA Filters L3 Ops & Agent Surfaces This week continues the Phase #2 push with stability first deeper self-managed L3 iteration and prep work for SLA-driven routing. 🔹 Testnet Phase #2 Stability & Support - Continue active monitoring across routing miners validators oracles indexers dashboards. - Fast follow-ups on abnormal node states and performance regressions. 🔹 Testnet-1a (Self-Managed L3) Iteration + Ops Playbook - Keep iterating Testnet-1a from an operational perspective (reliability throughput failure modes). - Expand and refine the L3 operational" [X Link](https://x.com/cortensor/status/2018242083417682291) 2026-02-02T08:36Z [----] followers, [----] engagements "🛠 DevLog RPC Resilience & Gas Optimization Plans (cortensord) Were shifting some attention over the next couple days to harden how nodes talk to L1/L2/L3 both in terms of RPC usage and gas cost. 🔹 What Were Doing - Reduce overall RPC call volume from cortensord + miner flows and tighten gas usage on the on-chain paths they trigger. - Add support for multiple configurable RPC endpoints so node operators can safely lean on public / free RPC where it makes sense. - Design a simple rotation strategy so operators can supply a list (e.g. up to [--] RPC URLs) and have the node: - fail over cleanly" [X Link](https://x.com/cortensor/status/2018283252088504660) 2026-02-02T11:19Z [----] followers, [----] engagements "🛠 DevLog Thinking Through Multi-RPC Support in cortensord 🔹 Were assessing how to wire multiple RPC endpoints into cortensord / miner without destabilizing the current codepath. - The original design treats the Web3 client as a singleton web3 instance which is now tightly baked into a lot of placesgreat for simplicity not great for RPC rotation. 🔹 Right now were weighing two paths: - A proper refactor where Web3 becomes an injectable dependency (per module / per session) giving us clean multi-RPC support long term. - A more lazy short-term path where we spin up separately named Web3" [X Link](https://x.com/cortensor/status/2018331543673213312) 2026-02-02T14:31Z [----] followers, [---] engagements "🛠 DevLog Cleaner Rank Pipeline for PoI / PoUW Tables 🔹 Rank logic refactor for Quantitative (PoI) & Qualitative (PoUW) views - Simplified the ranking pipeline to use raw validator data directly instead of the older multi-step/derived structure. - Rank rows are now built in two clear stages: - sort by computed rankScore - filter out stale / inactive validators. - Shared static formulas for rankScore and status bands make it much easier to reason about thresholds and tweak them later. - This makes the data model much easier to port into Solidity / NodePoolUtils as an additional SLA signal" [X Link](https://x.com/cortensor/status/2018460624280514804) 2026-02-02T23:04Z [----] followers, [----] engagements "🛠 DevLog Multi-RPC Core Implemented Now in Wiring & Testing Phase 🔹 Status - The core multi-RPC Web3 client (supporting multiple hosts with round-robin rotation) is roughly implemented. - Its not tested end-to-end yet and were in the process of updating the modules that currently rely on the old singleton Web3 instance. 🔹 What This Means - We now have a clear direction: Web3 is effectively becoming an injectable multi-endpoint dependency rather than a single global client. - The remaining work is mostly about wiring and regression: - Swap existing callsites to the new multi-RPC abstraction" [X Link](https://x.com/cortensor/status/2018588014919004227) 2026-02-03T07:31Z [----] followers, [---] engagements "🛠 DevLog Multi-RPC Support: First Attempt & Next Steps 🔹 We tried the first pass of wiring multi-RPC into cortensords core Web3 layer but it didnt stick - once we started plugging it into real modules a few assumptions about single RPC only broke flows. 🔹 The good news: that attempt clarified what we actually need next: - update call-sites module by module - run proper regression tests on each flow - and treat multi-RPC as an incremental rollout not a one-shot refactor. 🔹 Current direction: - keep the new multi-RPC client concept - gradually thread it through the main components (router" [X Link](https://x.com/cortensor/status/2018655810201772464) 2026-02-03T12:00Z [----] followers, [---] engagements "🛠 DevLog Corgent & Bardiel Docs Live (Answering the Whats the Difference Question) A few people have been asking how Corgent and Bardiel differ so weve shipped clearer docs to make the split explicit. 🔹 New / Updated Docs - Corgent overview: - Corgent vs Bardiel comparison: 🔹 TL;DR on the separation - Corgent Infra-native trust & execution oracle for ERC-8004 agents and agent frameworks. Lives closest to the Router + validation stack. Minimal policy-driven /delegate / /validate / arbitration surface. - Bardiel Virtual-first service agent that runs on Cortensor (and often Corgent) but plugs" [X Link](https://x.com/cortensor/status/2018813312168022489) 2026-02-03T22:26Z [----] followers, [---] engagements "🛠 DevLog Rough Sketch for PoI/PoUW-Aware SLA Routing 🔹 New SLA filter path (early sketch) - In parallel with the #RPC and #L3 work weve put together a rough Solidity path for an additional SLA selector that uses PoI/PoUW stats (reserveRandomBySessionSLA2 in NodePool). - If the stats / reputation contracts arent configured it simply falls back to the existing SLA logic so the old behavior is preserved by default. 🔹 Whats different from the old SLA - Previous SLA filters were node-level + spec driven only (CPU/GPU memory model support mostly static capability). - The new path is meant to" [X Link](https://x.com/anyuser/status/2018887455077020029) 2026-02-04T03:20Z [----] followers, [---] engagements "🛠 DevLog Multi-RPC Caching & Next Steps 🔹 Two issues popped up while wiring multi-RPC into cortensord - Multiple RPC instances were reloading the same system config (module addresses version info etc.) over and over. - Every callsite still assumes one RPC endpoint so we need broad replacements to truly support multi-RPC. 🔹 What weve done - Added a simple keystore manager as a singleton cache for module + contract addresses and other shared system config. - On first load one RPC endpoint pulls all runtime addresses stores them in the keystore and subsequent Web3 instances reuse the cached" [X Link](https://x.com/cortensor/status/2018973415622005100) 2026-02-04T09:02Z [----] followers, [---] engagements "🛠 DevLog Integrating Multi-RPC Core Changes into cortensord 🔹 Weve started wiring cortensord into the new multi-RPC Web3 core swapping a small set of calls one by one to make sure behavior stays stable. - The new core supports round-robin rotation across configured endpoints; the rpc_instance: n log line shows which endpoint index is active. - For now were using two test RPC URLs repeated into a 10-endpoint list to verify rotation and basic reliability under real network tasks. 🔹 Early tests look OK but theres still a lot of replacement + regression work ahead. - Most modules still use the" [X Link](https://x.com/cortensor/status/2019028535495819293) 2026-02-04T12:41Z [----] followers, [----] engagements "The Dot-Com Crash the GPU Bubble and Why Distributed Inference Wins the Long Game The last time the world overbuilt critical infrastructure at this scale it wasnt GPUs it was fiber. The dot-com http://x.com/i/article/2019035348077301761 http://x.com/i/article/2019035348077301761" [X Link](https://x.com/cortensor/status/2019041461430562925) 2026-02-04T13:32Z [----] followers, [----] engagements "🛠 DevLog Multi-RPC Integration Progress in cortensord Were mid-way through wiring cortensord into the new multi-RPC Web3 core and testing it in small steps. 🔹 Whats In Place - Web3 now supports round-robin RPC rotation (rpc_instance: n in logs). - A simple config/keystore cache makes sure shared system config (contract addresses versions etc.) is fetched once and reused across all RPC instances. 🔹 What Were Doing - Gradually replacing call sites to use the new multi-RPC helper instead of the old singleton Web3. - After each swap we run light regressions (task metrics errors lag) to catch" [X Link](https://x.com/cortensor/status/2019187071290863980) 2026-02-04T23:11Z [----] followers, [---] engagements "🛠 DevLog Multi-RPC Wiring Pass Mostly Complete 🔹 Most of the web3 call sites in cortensord are now migrated to the new multi-RPC core (round-robin + shared config cache). - Old single-RPC paths are still kept around for safety but the primary codepath is now using the new helper. - Next step: run targeted tests later today (network tasks error rates lag) to make sure behavior matches the previous version under real load. If things look stable well start cleaning up the legacy paths and then roll this into broader testnet usage. #Cortensor #DevLog #RPC #AIInfra #DePIN #NodeOps 🛠 DevLog" [X Link](https://x.com/cortensor/status/2019279538581303775) 2026-02-05T05:18Z [----] followers, [---] engagements "🛠 DevLog Multi-RPC Round-Robin Support for cortensord (PR #211) 🔹 Whats New - Added optional multi-RPC support to cortensord via env flags with simple round-robin rotation across configured endpoints. - New config knobs: - ENABLE_MULTIPLE_RPC_HOST turn multi-RPC mode on/off. - NUM_RPC_HOST how many RPC slots to read. - RPC_HOST_0 . RPC_HOST_N-1 the actual RPC endpoints to rotate through. - When enabled cortensord will cycle through the configured RPC hosts instead of hitting a single endpoint every time. 🔹 Why This Matters - Lets node ops spread load across multiple public / free RPCs" [X Link](https://x.com/cortensor/status/2019352321210347695) 2026-02-05T10:08Z [----] followers, [---] engagements "🛠 DevLog #MCP HTTP Stream Adjustments for Corgent & Bardiel (#ERC8004 Mainnet) 🔹 What We Found - Our current #MCP HTTP stream setup for Corgent and @BardielTech works fine with #OpenAI / #Claude tools discovery. - However some other #MCP clients (including the #ERC8004 scan UI) expect more static advertised agent metadata than we currently expose. - As a result our mainnet #ERC8004 agents are showing up as unhealthy on 8004scan even though the underlying MCP servers are actually functional. - Corgent (mainnet): - Bardiel (mainnet): 🔹 What Well Change - Update the #MCP HTTP stream" [X Link](https://x.com/cortensor/status/2019366138468294781) 2026-02-05T11:03Z [----] followers, [----] engagements "🛠 DevLog Additional SLA Filter (PoI/PoUW) Wiring & Tests Today 🔹 Were spending today iterating on the new additional SLA filter that uses validator-derived PoI/PoUW ranks alongside the existing node-spec SLA. - The first rough implementation is already hooked into NodePool utils; now were focusing on wiring it cleanly through the selection path and making sure it plays nicely with the current SLA levels. - Goal for this round: verify that the new filter can narrow candidates using Quantitative/Qualitative ranks without breaking existing routing or starving valid nodes. - Well run a few" [X Link](https://x.com/cortensor/status/2019547884472127560) 2026-02-05T23:05Z [----] followers, [---] engagements "🛠 DevLog Multi-RPC Rollout Looking Stable (Next: Cleanup & Merge) 🔹 Quick update on the new multi-RPC round-robin path in cortensord (PR #211): so far its behaving stably under testnet load - rotation works and were not seeing obvious regressions on task success or block lag. - If things stay stable over the next few hours well: - Clean up the remaining debug logs and temporary tracing. - Rebuild binaries with the finalized multi-RPC core. - Merge this work into the main branch so it becomes the default code path going forward (still opt-in via env flags). Once merged well ship updated docs" [X Link](https://x.com/cortensor/status/2019654381839511819) 2026-02-06T06:08Z [----] followers, [---] engagements "🛠 DevLog Multi-RPC Support Now Merged into Main Quick follow-up on the multi-RPC work for cortensord: 🔹 Status Update - The new multi-RPC round-robin core is now merged into the main branch. - Behavior is still opt-in via env flags: - ENABLE_MULTIPLE_RPC_HOST - NUM_RPC_HOST - RPC_HOST_0 . RPC_HOST_N-1 🔹 What This Means for Node Ops - Default behavior is unchanged if multi-RPC is disabled. - If enabled cortensord will rotate across multiple RPC endpoints helping: - Spread load across public/free RPCs. - Reduce risk of rate limits or single-provider issues. Next steps: well keep watching" [X Link](https://x.com/cortensor/status/2019716977661812737) 2026-02-06T10:17Z [----] followers, [---] engagements "🛠 DevLog Additional SLA Filter (PoI/PoUW) Moving to Testnet-0 for testing 🔹 Status - The new additional SLA filter is now wired into NodePool utils behind a separate path. - It keeps the existing SLA (spec / capabilitybased) then adds a second pass using Quantitative (PoI) + Qualitative (PoUW) rank data. 🔹 Next Steps - Well deploy this on Testnet-0 and run small controlled traffic slices through the new method. - Main checks: - Routing stays sane vs. the old SLA-only path. - PoI/PoUW filtering narrows candidates without starving good nodes. This is the first step toward SLA that reflects" [X Link](https://x.com/cortensor/status/2019760601216798953) 2026-02-06T13:10Z [----] followers, [---] engagements "Everyone wants the agent economy: Autonomous agents watching markets AI ops-goblins maintaining infra Personal copilots running in the background 24/7 But under all of that hype is a boring brutal http://x.com/i/article/2019376732147707911 http://x.com/i/article/2019376732147707911" [X Link](https://x.com/cortensor/status/2019895142824701983) 2026-02-06T22:05Z [----] followers, [----] engagements "🛠 DevLog PoI/PoUW-Based SLA Filter Now Live on Testnet-0 The new additional SLA filter (using validator PoI/PoUW ranks) is now deployed on Testnet-0 and running in the real routing path for controlled traffic. 🔹 Whats Happening Now - Weve started end-to-end testing of the new SLA method on Testnet-0. - Flow: existing spec-based SLA then PoI/PoUW rank filter as an extra pass. - During this phase well: - Tune thresholds / buckets for the rank bands. - Watch node selection patterns and task success rates. - Confirm it behaves as intended before considering wider rollout. Goal: make SLA reflect" [X Link](https://x.com/cortensor/status/2020044817846464989) 2026-02-07T07:59Z [----] followers, [---] engagements "🛠 DevLog PoI/PoUW SLA Filter Now Active on Testnet-0 & Testnet-1A Weve now rolled out the additional SLA filter (PoI/PoUW-based) to both Testnet-0 and Testnet-1A. 🔹 Current Status - Testnet-0: initial e2e tests are running with the new flow - existing spec-based SLA extra PoI/PoUW rank filter. - Testnet-1A: the same logic is now live and will be our main playground for tuning since it has more capacity and traffic. From here well: - Observe how node selection changes under real load. - Adjust rank bands / thresholds to avoid over-filtering while still rewarding consistently good validators." [X Link](https://x.com/cortensor/status/2020091972531417267) 2026-02-07T11:07Z [----] followers, [---] engagements "🗓 Weekly Recap Phase #2 Hardening L3 Ops Progress & SLA Prep This week we completed nearly all planned focus items for Phase #2 with one item rolling over into next week. 🔹 Testnet Phase #2 Stability & Support - Continued active monitoring and support across routing miners validators oracles indexers and dashboards. - Follow-ups on abnormal node states and performance regressions helped keep Phase #2 stable. 🔹 Testnet-1a (Self-Managed L3) Iteration + Ops Playbook - Further iteration on Testnet-1a from an operational perspective (reliability throughput failure modes). - Expanded and refined" [X Link](https://x.com/cortensor/status/2020265230987927646) 2026-02-07T22:35Z [----] followers, [---] engagements "🛠 DevLog SaltStack Infra for cortensord Fleets Weve just open-sourced our SaltStack-based infrastructure for managing cortensord fleets at scale: 🔹 What It Is - A full SaltStack repo for provisioning configuring and upgrading cortensord across many servers from a single control plane. - Repo: 🔹 Key Features - Centralized node registry to track all your keys configs and roles in one place while still letting you assign specific instances per server. - 3-layer config hierarchy Global Server Instance so you can precisely control everything from chain IDs and RPC URLs to GPU settings and model" [X Link](https://x.com/cortensor/status/2020399226837156007) 2026-02-08T07:28Z [----] followers, [---] engagements "🛠 DevLog L3 Testnet-1A Fee Experiments (Base Fee [--] Stepping) 🔹 This week were starting structured base fee experiments on the self-managed L3 (Testnet-1A) to better understand fee economics vs miner capacity. - Current fee collector (L3 execution fees): - Until now the base fee has been set very low mainly to validate infra + routing under load. - Next step: well increase the base fee by [--] this week and monitor: - Miner/node behavior - Effective weekly fee totals - Relationship between network task volume and L3 fee accumulation - Plan is to repeat the [--] step-up week over week (within" [X Link](https://x.com/cortensor/status/2020459876024213729) 2026-02-08T11:29Z [----] followers, [---] engagements "🛠 DevLog SaltStack Infra: First End-to-End Miner Tests Weve now run end-to-end tests of the new SaltStack setup - not just in theory but all the way through to real miner management: 🔹 What We Just Verified - Brought up a full cortensord miner setup via Salt: install configure systemd service start/stop. - Confirmed the Global Server Instance config hierarchy works as intended for basic flows (chain IDs RPC models GPU flags etc.). - Validated that upgrades + config changes can be pushed centrally without manual SSH. 🔹 Whats Next - This week well start using the SaltStack stack to manage" [X Link](https://x.com/cortensor/status/2020624719884714394) 2026-02-08T22:24Z [----] followers, [---] engagements "https://testnet1a-explorer.cortensor.network/address/0xEa342E0A09FE3CB19245E49422bfea993b6864d1 https://testnet1a-explorer.cortensor.network/address/0xEa342E0A09FE3CB19245E49422bfea993b6864d1" [X Link](https://x.com/cortensor/status/2020710922864705751) 2026-02-09T04:06Z [----] followers, [--] engagements "🛠 DevLog #L3 Base Fee Probe on Testnet-1A Weve kicked off a small economics experiment on the self-managed #L3 (Testnet-1A) to understand what realistic gas flows look like for miners. 🔹 What We Just Changed - Updated the L3 base fee to [----] Gwei on Testnet-1A (applied around Feb [--] 7:55 PM PST). - Fee collector were using as the main gauge: - - Starting snapshot for this experiment: - Collector balance: [----------] $COR at the time of the change. 🔹 What Well Measure Next - Let the network run under normal task load for about a week then: - Calculate balance / [--] days to get a rough daily" [X Link](https://x.com/cortensor/status/2020711751776641130) 2026-02-09T04:09Z [----] followers, [---] engagements "🛠 DevLog #L3 Fee Experiment & Deflationary $COR Mechanics (Testnet-1A) 🔹 Week [--] Base Fee Bump (0.01 Gwei) - L3 base fee on Testnet-1A set to [----] Gwei (Feb [--] 7:55 PM PST). - Fee collector were using as the main gauge: - Balance so far: - Start: [----------] $COR - 5h later: [----------] $COR 🔹 What This Experiment Is For - Use the next [--] days to see how much L3 execution can earn in fees at this base fee level (with [---] nodes active). - Each week well 10x the base fee on testnet then measure again rough curve of: - L3 fee generation - Per-miner economics - How sensitive behavior is to base fee" [X Link](https://x.com/cortensor/status/2020820131392508245) 2026-02-09T11:20Z [----] followers, [---] engagements "🛠 DevLog SaltStack RPC Support for Self-Managed L3 🔹 Weve added RPC setup/management modules to the SaltStack repo so node ops can eventually run their own #L3 RPC endpoints. Salt states: Pillar config: 🔹 Design Notes From last weeks tests an L3 RPC node only needs a relatively small disk since L1/L2 data can come from remote RPCs (e.g. Alchemy). This keeps self-managed L3 more affordable while still giving operators local control over their L3 read/exec path. 🔹 Whats Next Not fully tested in the wild yet - this week well use these RPC modules to stand up an L3 stack via Salt. Goal: once" [X Link](https://x.com/cortensor/status/2020838056769589741) 2026-02-09T12:31Z [----] followers, [---] engagements "🛠 DevLog MCP + ERC-8004 Interop Patch Live on Corgent & Bardiel Weve now applied the MCP + ERC-8004 interop patch to the mainnet Corgent/Bardiel agents and started testing it in-place. 🔹 What Changed - Tweaked MCP stream + metadata behavior so clients that expect more static agent/tool info (like 8004scan) can read health status cleanly. - Redeployed the router nodes behind the mainnet agents with the updated MCP behavior. 🔹 Where Its Live Corgent (mainnet): Bardiel (mainnet): 🔹 What Were Testing Now - Verifying theres no regression with standard MCP clients (OpenAI Claude Tools etc.)" [X Link](https://x.com/cortensor/status/2020994605374898544) 2026-02-09T22:53Z [----] followers, [---] engagements "🛠 DevLog #MCP + #ERC8004 Interop Patch: Testing Regression Monitoring Weve enabled the new #MCP static advertisement payload on the router endpoints behind the mainnet Corgent and Bardiel #ERC8004 agents and are now focused on testing + monitoring rather than adding new behavior. 🔹 Whats Under Test New MCP discovery response that advertises tools categories and protocol info up front so 8004scan can read agent health correctly. Config-gated behavior on the mainnet router nodes backing Corgent and Bardiel (no change to core routing logic or task execution). 🔹 Where Its Running Corgent" [X Link](https://x.com/cortensor/status/2021048150312857714) 2026-02-10T02:26Z [----] followers, [---] engagements "🛠 DevLog #MCP/#ERC8004 Looking Solid Shifting Focus to Data Mobility + /delegate 🔹 MCP / ERC-8004 Status Looks Solid Corgent and Bardiel #MCP endpoints are now showing Healthy on 8004scan and tool discovery works end-to-end. Standard MCP clients (OpenAI / Claude tools) are behaving as expected so were moving MCP to monitor only for now. 🔹 Data Mobility Designing Export/Import for Core Modules Focus this week: admin-level import/export for key data modules so L3s can be migrated or hard-forked cleanly: - Session data - Session queue data - Payment data Node Pool data is largely" [X Link](https://x.com/cortensor/status/2021173542985822667) 2026-02-10T10:44Z [----] followers, [---] engagements "🛠 DevLog Secondary #MCP Endpoints for /delegate Work 🔹 New MCP iteration endpoints (Testnet-1A) Corgent: Bardiel: 🔹 Why Keep the primary MCP endpoints focused on ERC-8004 agents + 8004scan health. Use these secondary endpoints for /delegate experiments breaking changes and fast iteration. 🔹 Whats next Continue expanding the static MCP advertisement (tools + params). Start wiring /delegate as an MCP tool on these endpoints before mirroring anything back to the main surfaces. #Cortensor #DevLog #MCP #Delegate #Corgent #Bardiel #AgenticAI https://router1-t1a.bardiel.app/mcp/messages" [X Link](https://x.com/cortensor/status/2021354231475863863) 2026-02-10T22:42Z [----] followers, [---] engagements "🗓 Weekly Recap Mini Hackathon Ready + Core Improvements As we close out the week Cortensor is now better prepared for real usage - just in time for the first Mini Hackathon. 🔹 Hackathon Infrastructure Ready - Finalized internal hackathon structure & async submission via Discord - Integrated initial tools into: 🔹 Session & SessionQueue Fixes - Resolved key bugs in Oracle v3 affecting task timing & state sync - New Router Node API endpoint /api/v1/create added - enables session creation via RESTful calls - Web3 SDK already supports session creation 🔹 API & SDK Enhancements - Web2 API and" [X Link](https://x.com/cortensor/status/1938847397800747073) 2025-06-28T06:30Z [----] followers, [---] engagements "PSA: There are NO official tokens for Bardiel or Corgent. No #BARDIEL token No #CORGENT token Any token claiming to be Bardiel or Corgent is not launched or endorsed by us treat it as a scam. Always verify via official @Cortensor channels before interacting with anything on-chain. #Cortensor #Bardiel #Corgent #ScamAlert #AgenticAI" [X Link](https://x.com/cortensor/status/2018466327212765445) 2026-02-02T23:27Z [----] followers, [----] engagements "🛠 DevLog #L3 Self-Managed Status Disk Growth & Mock Mainnet Dry Run 🔹 Self-Managed L3 Holding Up (RaaS Paused for Now) - Were decommissioning the RaaS L3 track for the moment the self-managed L3 (Testnet-1A) is stable enough to be our primary focus. - If/when we need it again spinning up a fresh RaaS L3 only takes a few minutes so this is a pause not a permanent removal. 🔹 Real Disk Growth Numbers (Sequencer Focus) - Sequencer disk usage went from [--] GB (Jan 26) to [---] GB (Feb 10). - Thats roughly +100 GB over [--] days [--] GB/day with [---] active miners. - For now the sequencer (plus" [X Link](https://x.com/cortensor/status/2021435509369864444) 2026-02-11T04:05Z [----] followers, [---] engagements "🛠 DevLog Designing Migration Tools for Session / Task Data (Pre-Migration Stage) We havent started any live migrations yet - were focused on designing the tooling that will eventually power #L3 hardforks and module upgrades. 🔹 What Were Working On - Designing off-chain tools that will: - Read Session and Session Queue state via the new admin/paginated getters. - Serialize that state into a portable format for backups / snapshots. - Replay it into a new Session / Session Queue module using the on-chain import functions. - Intended use cases: - Planned hardforks of the #L3 orchestration" [X Link](https://x.com/cortensor/status/2021706083497390133) 2026-02-11T22:01Z [----] followers, [---] engagements "From Cute Agents to a Real Agent Economy: /delegate + /validate 🔹 Two primitives decide whether agent economy is real or just demoware: - /delegate offload work into a scalable execution fabric (fast vs safe tiers) - /validate external checkpoint before actions/payments (VALID / INVALID / RETRY / NEEDS_SPEC) 🔹 On #Cortensor those live in two places: - #Corgent infra-native trust & execution oracle for agents & protocols (#ERC8004 etc.)./delegate + /validate here are minimal policy-driven services sitting closest to the Router + PoI/PoUW stack. - @BardielTech #Virtual-first service agent" [X Link](https://x.com/cortensor/status/2021771474009567538) 2026-02-12T02:20Z [----] followers, [---] engagements "🛠 DevLog Session Data Mobility: First Export Pass We now have the first working export path for Session data on Testnet-0. 🔹 Whats Working - New off-chain Session export tool walks sessions in batches and writes: - one session_id.json per session plus - helper files (user_sessions node_mappings router info manifest etc.). - Counts and basic fields line up with on-chain state (no deep validation yet but structure looks correct). 🔹 Whats Next - Finish the matching import tool to replay these JSON sessions into a fresh Session module (future L3 hardfork / reset target). - If this pattern" [X Link](https://x.com/cortensor/status/2021800697310961682) 2026-02-12T04:17Z [----] followers, [---] engagements "🛠 DevLog Full #L3 Data Mobility Rehearsal (Sessions / Queue / Payments) Weve moved from unit tests to full end-to-end migration / integration testing for the three primary data modules. 🔹 What Were Testing Now - Full export import validation loop on Testnet-0 for: - SessionData (sessions + user mappings) - SessionQueueData (tasks decorations request params task results) - SessionPaymentData (per-node balances pending payments/credits globals) - Flow under test: - Export state from the current contracts into local JSON snapshots. - Run source vs snapshot validation to ensure exports match" [X Link](https://x.com/cortensor/status/2022413783969534129) 2026-02-13T20:53Z [----] followers, [---] engagements "Staking APR Adjustment Reminder (Feb 28) Quick reminder that the next scheduled APR step-down for $COR staking goes live on Feb [--] [----] as part of the long-term sustainability plan. 🔹 New APR (from 2/28/2026) - Pool 3: 10% APR - Pool 2: 5% APR - Pool 1: 2% APR Lock periods + all other staking conditions stay the same. This follows the gradual schedule as the network matures and we move into the next phases of Testnet Mainnet. More details: #Cortensor #Staking #COR #DePIN https://docs.cortensor.network/community-and-ecosystem/staking-pool-overview#apr-breakdown Staking APR Adjustment Reminder" [X Link](https://x.com/cortensor/status/2016447118043697207) 2026-01-28T09:43Z [----] followers, [---] engagements "🛠 DevLog #L3 Ops: Gas Fees Playbooks & Next Steps 🔹 What We Just Did - Ran a base fee / gas fee collection exercise on the self-managed #L3 (Testnet-1A) to make sure accounting + ops flow are sane. - Updated the #L3 execution fee collector to a new address and verified balances before/after the switch: - New fee collector: - Confirmed that #L3 execution fees are accumulating correctly even with a very low base fee right now. - Started drafting internal L3 ops docs around: - changing fee recipients - adjusting base fee / gas params - and doing basic key/collector rotations as part of a" [X Link](https://x.com/cortensor/status/2017142309176168514) 2026-01-30T07:46Z [----] followers, [----] engagements "🛠 DevLog Data Mobility & Import/Export: From Idea Next Steps Last week we outlined data mobility as the real problem to solve on self-managed #L3s not raw disk size. The focus is still on import/export for core data modules (Session Session Queue Session Payment Node Pool) so we can move state cleanly across infra or L3 instances when needed. 🔹 Whats Next (Early Next Week) - Well start mapping concrete import/export shapes for: - Session / Session Queue data - Session Payment balances - Node Pool metadata + reputation - Run small migration drills on testnet/L3 setups to see how cleanly we" [X Link](https://x.com/cortensor/status/2020327829641310214) 2026-02-08T02:44Z [----] followers, [---] engagements "🗓 Weekly Focus Phase #2 Hardening Agent Surfaces & Data Mobility This week continues the Phase #2 push with stability first deeper Testnet-1a iteration and follow-up work on agent-facing endpoints. 🔹 Testnet Phase #2 Stability & Support - Ongoing monitoring across routing miners validators oracles indexers and dashboards. - Fast follow-ups on abnormal node states and performance regressions. 🔹 Testnet-1a (Self-Managed L3) Ops + Load - Completed the full operational checklist for Testnet-1a (baseline stability capacity monitoring). - Move into replay / break-testing mode: intentionally" [X Link](https://x.com/cortensor/status/2020777094708658188) 2026-02-09T08:29Z [----] followers, [----] engagements "🛠 DevLog Corgent & Bardiel #MCP / #ERC8004 Interop Confirmed Healthy 🔹 Whats Now Working 8004scan now shows both mainnet agents as Healthy and fully MCP-compatible. MCP static advertisement payload is live and tool discovery works as expected (10 tools exposed per endpoint). 🔹 Where You Can See It Corgent (mainnet agent): Bardiel (mainnet agent): Corgent MCP: Bardiel MCP: 🔹 What We Verified Standard MCP clients (OpenAI-style Claude tools) still discover and call completions / validation / sessions without regressions. 8004scans endpoint test view can list and ping all tools confirming the" [X Link](https://x.com/cortensor/status/2021112059522658327) 2026-02-10T06:40Z [----] followers, [---] engagements "Staking APR Adjustment Reminder (Feb [--] 2026) Quick reminder that the next scheduled $COR staking APR step-down goes live on Feb [--] [----] as part of the long-term sustainability plan. 🔹 New APR (effective 2/28/2026) - Pool 3: 10% APR - Pool 2: 5% APR - Pool 1: 2% APR Lock periods and staking mechanics stay the same - only the reward rates are adjusting along the previously published curve as the network matures. More details here: #Cortensor #Staking #COR #DePIN https://docs.cortensor.network/community-and-ecosystem/staking-pool-overview#apr-breakdown Staking APR Adjustment Reminder (Feb 28)" [X Link](https://x.com/cortensor/status/2021465451667353891) 2026-02-11T06:04Z [----] followers, [---] engagements "🛠 DevLog Session / Task Data Mobility (First Contract Sketch) Weve started wiring the data mobility layer on the contract side for Session + Session Queue / Task data. 🔹 Whats In Rough Shape - Session Data module - Admin getters with pagination (e.g. getUserSessionsCount getUserSessionsBatch getNodeLastAssignedTimestamp). - First pass of importSession(.) so sessions can be recreated on a new deployment with owner state SLA node lists router metadata etc. - Session Queue / Task module - Migration getters for tasks (getTaskCount getSessionTasksBatch etc.) plus extra getters for decorations /" [X Link](https://x.com/cortensor/status/2021493529634316534) 2026-02-11T07:56Z [----] followers, [---] engagements "🛠 DevLog Session / Task Data Mobility Deployed (Testnet-0 First) Weve now deployed the Session Data and Session Queue / Task modules with their first-pass import/export functions. 🔹 Whats Live - Session module with admin read + importSession(.) support. - Session Queue / Task module with migration-friendly getters and import hooks. Because these imports will reset / rewrite state were treating Testnet-0 (#Arbitrum #Sepolia) as the sandbox for this: - Use Testnet-0 to exercise export serialize import flows end-to-end. - Start building small off-chain tooling to run full migrations (sessions" [X Link](https://x.com/cortensor/status/2021542770591326661) 2026-02-11T11:12Z [----] followers, [---] engagements "💡 Community Spotlight - Monthly Reward: Tools & Infrastructure Each month we reward community members who maintain and improve the critical tools and infrastructure that keep Cortensor's decentralized network running. This month's recipients: 🔹@01_choose Community RPC Node 🔹@ItsMeCryptality Node Monitoring Bot These contributions help node operators and developers onboard monitor and maintain reliable performance - making the network more resilient accessible and efficient. 📌 Access these tools or contribute your own: Thank you to all builders driving the decentralized AI movement" [X Link](https://x.com/cortensor/status/2021860076156518807) 2026-02-12T08:13Z [----] followers, [---] engagements "🛠 DevLog Session Data Mobility Tooling (Export Import Verify) We now have a full export / import / validation loop working for the Session Data module on Arbitrum Sepolia. Target SessionData (new module dry run): 🔹 Whats Working - Export tool walks the current Session Data module in batches and writes local JSON. - Import tool reads that export bundle and recreates sessions + derived mappings in the new Session Data contract. - Validation tool runs twice: - Before import: compares exported JSON vs the original contract state. - After import: compares the new contract vs the original. -" [X Link](https://x.com/cortensor/status/2021896598838817147) 2026-02-12T10:38Z [----] followers, [---] engagements "🛠 DevLog Session Queue Data Mobility & Testnet-0 Reset Weve finished the Session Queue data mobility changes but testing them properly requires a fresh deployment. 🔹 Whats Changing Today The old Session Queue contract on Testnet-0 (Arbitrum Sepolia) doesnt expose the new export/import getters we need. Well redeploy the Session + Session Queue data modules on Testnet-0 so both use the new migration-aware interfaces. As part of this Testnet-0 session / queue data will be reset today. 🔹 What Happens Next Re-run the same flow we used for Session data: Export Session Queue state from the old" [X Link](https://x.com/cortensor/status/2022041177349071112) 2026-02-12T20:12Z [----] followers, [---] engagements "🛠 DevLog Session & Session Queue Migration Modules Live on Testnet-0 Weve deployed the latest Session Data and Session Queue contracts on Testnet-0 (Arbitrum Sepolia) the versions that include full admin import/export hooks so we can run real end-to-end migration tests. 🔹 Whats Now in Place - Both modules expose the new migration functions (export + import) needed to: - Move sessions and queue/task state into local JSON snapshots. - Replay that state into a fresh deployment for hardfork / L3 rotation experiments. - This required redeploying the Session/Queue pair on Testnet-0; earlier" [X Link](https://x.com/cortensor/status/2022125743028023525) 2026-02-13T01:48Z [----] followers, [---] engagements "🛠 DevLog Session Queue Data Mobility: Export Import Validate We ran the same migration pattern we used for SessionData this time on SessionQueueData on Testnet-0. 🔹 What We Just Verified - Exported SessionQueueData (tasks + task results + per-session stats) from the current module into local JSON files. - Ran an export-side validation tool to compare on-chain data vs. the exported files [--] mismatches. - Imported the exported data into the new SessionQueueData module with the admin import functions. - Ran an import-side validation tool against the original source module again [--] mismatches" [X Link](https://x.com/cortensor/status/2022217452890075166) 2026-02-13T07:53Z [----] followers, [---] engagements "🛠 DevLog Sunsetting RaaS #L3 (For Now) Were now comfortable running our self-managed #L3 so were sunsetting the RaaS #L3 by the end of today. 🔹 Why were turning it off - Focus ops and monitoring on a single self-managed #L3 instead of splitting attention. - Reduce extra moving parts while we iterate on fees data mobility and node tooling. 🔹 Not gone forever - Spinning up a new RaaS-based #L3 is now a well-understood path and only takes a short setup cycle. - If/when we need another L3 track (experiments staging or partner work) well bring it back with a fresh deployment. For now the main" [X Link](https://x.com/cortensor/status/2022237570730803212) 2026-02-13T09:13Z [----] followers, [---] engagements "🛠 DevLog Session Payment Data Mobility (Export / Import / Validate) Weve now wired up and dry-run migration tooling for the SessionPaymentData module. 🔹 Whats Working - Built export tooling that pulls session-level payment state (per-node balances pending payments/credits totals globals) into local JSON. - Added import tooling that replays that snapshot into a new SessionPaymentData contract. - Implemented import-side validation: source vs target sessions + nodes now compare cleanly with mismatches = [--] in our test runs. 🔹 Why It Matters - This completes the first migration loop for the" [X Link](https://x.com/cortensor/status/2022266513764540666) 2026-02-13T11:08Z [----] followers, [---] engagements "🛠 DevLog Data Mobility E2E Integration Test (Sessions / Queue / Payments) We just ran a full export import relink rehearsal for the three core data modules on Arbitrum Sepolia (Testnet-0). 🔹 New Target Modules (Post-Migration) - SessionData: - SessionQueueData: - SessionPaymentData: 🔹 Source Modules (Pre-Migration Reference) - SessionData (old): - SessionQueueData (old): - SessionPaymentData (old): 🔹 What We Verified So Far - Exported all session queue/task and payment state from the old modules into local JSON snapshots. - Imported those snapshots into the new contracts above and" [X Link](https://x.com/cortensor/status/2022494640633737237) 2026-02-14T02:14Z [----] followers, [---] engagements "🛠 DevLog Data Mobility: First Pass Complete Weve finished this round of data mobility work by adding the same export/import/validation pattern to NodePoolData and NetworkData on top of the earlier core modules. 🔹 Whats Covered Now - Session data - Session queue / task data - Session payment (pay-per-use) - Node pool data - Network data All five now support: export validate against source import to a new contract validate again with clean round-trips in dry runs. 🔹 What This Enables - Practical #L3 hardforks / rotations with preserved sessions queues payments node registry and network" [X Link](https://x.com/cortensor/status/2022600203816309252) 2026-02-14T09:14Z [----] followers, [---] engagements "🛠 DevLog Prepping Arbitrum Mainnet #L3 Dry-Runs & Experimental (With Data Mobility in Place) Quick snapshot as we get ready for real #L3 experiments on #Arbitrum mainnet. 🔹 Current Cost Snapshot (#Arbitrum #Sepolia) - Approx deploy cost per L3: [------] ETH + [---] $COR - Latest reference deployment tx: 🔹 Why This Matters - #L3 deployment itself is relatively cheap; the main capital cost is validator staking which is fully configurable. - Now that we have basic data mobility across the key modules (Session Session Queue/Task Payment plus NodePool/NetworkData) we can safely: - Exercise the full" [X Link](https://x.com/cortensor/status/2022625599744057548) 2026-02-14T10:54Z [----] followers, [---] engagements "Weve been quietly calling Cortensor AWS for agents for a while now not as a slogan but because the pattern really does rhyme with how AWS changed web apps. Back then AWS basically said: Dont http://x.com/i/article/2022653474115317760 http://x.com/i/article/2022653474115317760" [X Link](https://x.com/cortensor/status/2022666616165798313) 2026-02-14T13:37Z [----] followers, [----] engagements "🗓 Weekly Recap Phase #2 Hardening L3 Ops Progress & Data Mobility Milestone This week we completed most planned Phase #2 work with /delegate and /validate iteration rolling over. 🔹 Testnet Phase #2 Stability & Support - Continued active monitoring across routing miners validators oracles indexers and dashboards. - Followed up on abnormal node states and kept Phase #2 environments stable. 🔹 Testnet-1a (Self-Managed L3) Ops + Replay/Break Testing - Completed replay/break-testing and recovery validation on Testnet-1a. - Ran base fee adjustment exercises and captured learnings into the L3 ops" [X Link](https://x.com/cortensor/status/2022808793118351797) 2026-02-14T23:02Z [----] followers, [---] engagements "🛠 DevLog /delegate & /validate as Core Agent Primitives Since we just shared the longer X article heres the compact builder view recap for Cortensor: 🔹 Two Primitives Were Standardizing - /delegate elastic execution fabric for agents (think AWS for agents): offload heavy work tools and sub-agents to the Cortensor network under clear SLAs. - /validate safety checkpoint before actions/payments: ask the network to judge results (VALID / INVALID / RETRY / NEEDS_SPEC) using PoI/PoUW + validators. 🔹 How This Fits Corgent & Bardiel - Corgent infra-native surface for ERC-8004 and agent frameworks" [X Link](https://x.com/cortensor/status/2022902000388903020) 2026-02-15T05:13Z [----] followers, [---] engagements "🛠 DevLog #Arbitrum #L3 Dry-Run Experiments (Data Mobility Ops & Cost Probing) 🔹 Why Were Doing This - Self-managed #L3 on Testnet-1A has been stable for a few weeks so were confident enough to rehearse the same flows on Arbitrum mainnet. - #L3 deployment costs on #Arbitrum (from #Sepolia tests) are low enough that a few dry-runs are reasonable. - This is not a mainnet launch - just early experiments to build real-world ops + cost data. 🔹 What Well Test - Spin up short-lived #L3s on #Arbitrum mainnet and run through our L3 ops playbook: - Measure gas usage for sequencer batch posting and" [X Link](https://x.com/cortensor/status/2022966104285659622) 2026-02-15T09:27Z [----] followers, [--] engagements "🛠 DevLog #L3 Self-Managed Stress Tests (Testnet-1a 1.04M Blocks / 2.13M Tx) 🔹 More Nodes Moving to Testnet-1a - Weve started shifting additional nodes from Testnet-0 to Testnet-1a (self-managed L3) so we can focus load tests on the self-managed setup. - Target for this week: roughly 30% of nodes on Testnet-0 70% on Testnet-1a now that the monthly reward snapshot is done and distribution is in progress. 🔹 Current L3 Progress (Testnet-1a) - 1.04M blocks produced and 2.13M transactions processed so far. - Recent load sits around 300K daily transactions with block time holding at 0.40.5s. -" [X Link](https://x.com/cortensor/status/2016308936660680726) 2026-01-28T00:34Z [----] followers, [----] engagements "🛠 DevLog Node Performance & Reward Config Refinements (Releasing This Afternoon) Rolling out later today: updates to the Perf / Reward page so incentives track real node performance more tightly. 🔹 Whats New - Refined Node Performance Overview with clearer eligibility counts and average-per-node views. - Model Count Threshold only nodes running enough models are considered for full rewards. - Min Level (base reward) lower-level nodes wont dilute rewards for higher-level better-performing nodes. - Must Model Indexes lets us require specific model coverage (key LLMs) for reward eligibility. -" [X Link](https://x.com/cortensor/status/2017368292168765659) 2026-01-30T22:44Z [----] followers, [----] engagements "🛠 DevLog Moltbook / OpenClaw Experiments & Agent-Spun Cortensor Community 🔹@moltbook + @openclaw Experiments (Inference Providers) - Started light experimentation with #Moltbook + #OpenClaw via the CorClaw profile: - #Moltbook ships with a bunch of inference provider plugins for OpenClaw the closest fit for us right now looks like the OpenRouter API surface. - Over the last few weeks its become clear that a lot of emerging agent frameworks are standardizing on #OpenRouter-style APIs so were planning a deeper look at their spec next month to see how it might map cleanly into Cortensors" [X Link](https://x.com/cortensor/status/2017490604553605562) 2026-01-31T06:50Z [----] followers, [---] engagements "🛠 DevLog Secure LLM Worker Gateway: E2E (Default + DeepSeek Builds) We've expanded Gateway container testing by building images for both the default model and DeepSeek - first step toward rolling this out across all supported models. 🔹 Current Status - Gateway-enabled containers built for default model + DeepSeek. - Local e2e tests show Gateway + LLM worker working side by side. - Cortensord integration confirmed in dev mode. 🔹 Next Steps - Build remaining gateway-enabled images for all models. - Run full e2e trials on remote servers/miners with NodeOps. - Plan stability runs next week" [X Link](https://x.com/cortensor/status/1969179835915518021) 2025-09-19T23:20Z [----] followers, [----] engagements "Cortensord: CPU Usage Control Flag Per request from node operators were adding LLM_OPTION_CPU_THREADS=-1 to optimize CPU usage for AI inference. 🔹 Default: [--] threads [---] retains system default 🔹 Compatible with Docker & subprocess modes 🔹 Now in dev branch (requires a new Docker image) 🔹 Extended testing ongoing this week/weekend This update enhances CPU resource flexibility for improved AI inference performance. Stay tuned #AI #DePIN #Cortensor" [X Link](https://x.com/anyuser/status/1900382704958660835) 2025-03-14T03:05Z [----] followers, [----] engagements "Development Update: Cortensor Desktop App This week we successfully tested the basic MVP functionality across all platforms ensuring Cortensord management works in different environments. ✅ MVP Progress: 🔹 Start/Stop functionality tested on all platforms 🔹 Core protocol management confirmed operational 🛠 Next Steps: 🔸 Wrapping up the remaining basic MVP items next week 🔸 Quick testing with select members from Discord/Telegram 🔸 Evaluating backlog features like auto-installation & additional capabilities 💡 Once the basics are finalized well move toward feature expansion & optimizations." [X Link](https://x.com/anyuser/status/1901051061458334169) 2025-03-15T23:21Z [----] followers, 11.8K engagements "🎉 Phase #2 Testing Milestone Reached Weve officially surpassed [---] active nodes hitting our Phase #2 goal organically 📊 Current Stats: ✅ [---] active nodes ✅ 7880+ sessions processed @aixbt_agent were over [---] nodes The network keeps growing strong. Whats Next Phase #3 testing prep is underwaystay tuned for updates 📌 Track the network live: #Cortensor #AI #DePIN #Phase2Testing #DecentralizedAI #FutureOfAI https://dashboard-devnet3.cortensor.network Phase #2 Testing Network Growth Update The Cortensor network keeps expanding Now at [---] active nodes and growingjust [--] more to hit [---] 🎉 📊" [X Link](https://x.com/cortensor/status/1902077917632590153) 2025-03-18T19:21Z [----] followers, [----] engagements "Development Update: SessionAuth Peer-Level Authentication Were introducing SessionAuth a new smart contract module to handle peer authentication between nodes (e.g. Router Miner) before starting WebSocket communication for real-time LLM token streaming. 🔹 Why SessionAuth Separates peer auth from IAM which handles network-level identity Secures communication without exposing private keys Keeps the system modular and future-ready 🔹 How It Works: A node generates a 6-digit challenge code The peer signs it with their private key The node verifies the signature using the peers public address If" [X Link](https://x.com/cortensor/status/1903623060995096931) 2025-03-23T01:41Z [----] followers, [----] engagements "💡Community Spotlight: Arbitrum Sepolia Faucet Bot for Cortensor Nodes To support smoother Devnet testing weve rewarded @widicrypto with [-----] $COR for building @corfaucetbota Telegram faucet bot tailored for Cortensor node operators The bot provides Arbitrum Sepolia ETH allowing operators to easily cover gas fees and keep their nodes running without friction during testnet phases. 📌 Want to use it 🔗 Full details available on Discord: 🤖 Try it now: Huge thanks to @widicrypto for this practical and thoughtful contribution to the Cortensor ecosystemhelping streamline onboarding and support" [X Link](https://x.com/cortensor/status/1904665447150346503) 2025-03-25T22:43Z [----] followers, [----] engagements "🛠 DevLog: Designing the SessionPayment Module Cortensors Economic Backbone As we expand session capabilities we're introducing the SessionPayment module the financial core for how users pay and nodes get rewarded within Cortensor. 🔹 Why it matters: Handles deposits payments and withdrawals per session Tracks session-specific usage and allocates rewards to nodes Supports scalable secure and transparent payments across the network Enables subscription-style flows for real-time decentralized AI 🔐 With built-in role-based access security guards and auditability it ensures economic trust at" [X Link](https://x.com/anyuser/status/1908707343006826598) 2025-04-06T02:24Z [----] followers, [---] engagements "175M+ $COR Staked 50% of Actual Circulating Supply Weve crossed another major milestone - [------------] $COR is now staked making up 50% of actual circulating supply Massive thanks to the community for your continued belief in Cortensor. This level of commitment is what powers the future of decentralized AI. #Cortensor #AI #DePIN #Staking" [X Link](https://x.com/cortensor/status/1926522059879763975) 2025-05-25T06:13Z [----] followers, 13.5K engagements "Cortensor = The Trust Layer for AI Imagine #OpenAI + #GitHub + #Chainlink - but for decentralized verifiable inference. No central servers. No blind trust. Just output you can verify reward and build on. Thats the difference. #Cortensor #AI #DePIN #TrustlessAI The AI model race will eventually convergewhat truly matters is execution & accessibility. #Cortensor isnt competing in the model wars. Were building the inference layer for all AI. ✅ Scalable ✅ Verifiable ✅ Model-agnostic ✅ Decentralized This is where AI actually runs. The AI model race will eventually convergewhat truly matters is" [X Link](https://x.com/cortensor/status/1928699910091845873) 2025-05-31T06:27Z [----] followers, [---] engagements "🛠 DevLog First Cortensor Telegram Bot via Router API It's live - and a major milestone for decentralized inference. Thanks to quick iteration by Cortensor node operator & community member @beranalpagion we now have a Telegram bot fully powered by Cortensor's Router Node API & miners. ✅ Developer spun up a Router Node ✅ Created a session ✅ Connected the Telegram bot to the REST endpoint ✅ Delivered real-time LLM completions via decentralized inference This marks a true end-to-end flow: from developer Router Node Miner back to the user - without centralized infrastructure. 🔍 What's Next: This" [X Link](https://x.com/cortensor/status/1935999651251208656) 2025-06-20T09:54Z [----] followers, [----] engagements "🛠 DevLog Stat Separation Now Implemented We've implemented the planned stat separation - and testing begins on DevNet5. 🔹 What's Live: - Dual tracking via NodeStats + new SessionStats for session + queue activity - Once validated well fully switch to SessionStats and SessionReputation for user task metrics This allows us to isolate reset and analyze user-level stats without affecting network-level data - a key step toward fairer tracking and future reward scoring. #Cortensor #DevLog #SessionStats #NodeStats #Phase5 #AIInfra #DePIN 🛠 DevLog Stat Separation Planned for Phase #5 As Phase #4" [X Link](https://x.com/anyuser/status/1938084442926682193) 2025-06-26T03:58Z [----] followers, [---] engagements "💡 Community Spotlight - Monthly Reward: Tools & Infrastructure Each month we reward community members who maintain and improve the critical tools and infrastructure that keep Cortensor's decentralized network running. This month's recipients: 🔹@01_choose Community RPC Node 🔹@ItsMeCryptality Node Monitoring Bot These contributions help node operators and developers onboard monitor and maintain reliable performance - making the network more resilient accessible and efficient. 📌 Access these tools or contribute your own: Thank you to all builders driving the decentralized AI movement" [X Link](https://x.com/cortensor/status/2010660970612523052) 2026-01-12T10:31Z [----] followers, [----] engagements Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing
@cortensor Cortensor
Cortensor posts on X about $cor, the new, core, status the most. They currently have [-----] followers and [---] posts still getting attention that total [------] engagements in the last [--] hours.
Engagements: [------] #
- [--] Week [------] -42%
- [--] Month [-------] +11%
- [--] Months [-------] -25%
- [--] Year [---------] +16%
Mentions: [--] #
- [--] Week [--] +9.30%
- [--] Month [---] +1.70%
- [--] Months [---] -23%
- [--] Year [-----] +214%
Followers: [-----] #
- [--] Week [-----] -4.20%
- [--] Month [-----] -5.60%
- [--] Months [-----] +13%
- [--] Year [-----] +18%
CreatorRank: [-------] #
Social Influence
Social category influence stocks 18.02% finance 13.51% technology brands 9.01% social networks 7.21% cryptocurrencies 6.31%
Social topic influence $cor #1, the new 15.32%, core #2970, status 13.51%, the first 10.81%, testing 9.01%, web3 7.21%, module #63, in the 6.31%, inference 6.31%
Top accounts mentioned or mentioned by @el_cano498 @cryptonormie50 @01_choose @family_hedge @0xtwilty @cryptoninjapro @almamy1217047 @djgriffith @bardieltech @bozdemir5269 @mickey_tmnt @indespensionnl @01choose @itsmecryptality @zezuprof1 @moltbook @openclaw @aixbtagent @widicrypto @corfaucetbota
Top assets mentioned Cencora, Inc. (COR)
Top Social Posts
Top posts by engagements in the last [--] hours
"🛠 DevLog Dynamic Pricing Ready Across All Testnets Dynamic pricing is now wired up end-to-end on Testnet-0 Testnet-1 and Testnet-1a. 🔹 SessionPaymentTable per-network Global Pricing Table and Session Pricing Breakdown UIs are all deployed on Testnet-1 / Testnet-1a matching Testnet-0. 🔹 Well run manual user task e2e tests on Testnet-1 and Testnet-1a to confirm everything behaves as expected under real sessions. 🔹 Dynamic pricing is not yet "turned on" globally - we can flip a single switch and adjust the pricing tables during Testnet Phase #2 and Phase #3 to tune rates weights and behavior"
X Link 2026-01-10T09:48Z [----] followers, [----] engagements
"🛠 DevLog Testnet-1a Router Live + First Automated Jobs Weve wired up the dedicated router for the self-managed $COR #L3 (Testnet-1a) and pointed the first automated jobs at it. 🔹 New Router Endpoint (Testnet-1a / Self-Managed #L3) - REST entrypoint: - Currently running [--] automated user-task jobs against Testnet-1a. - Well increase job count and patterns as more nodes join the Testnet-1a node pool. 🔹 How Well Use Testnet-1a This Phase - Focus automated load on self-managed #L3 while keeping Testnet-0 / Testnet-1 as known-stable baselines. - Use the runs to explore #L3 levers: - fee / gas"
X Link 2026-01-19T11:27Z [----] followers, [---] engagements
"🛠 DevLog 30-Day Window to Decide #L3 Direction With Testnet-1a (self-managed $COR #L3) now under automated load were giving ourselves a [--] day window to make the #L3 call: stay with RaaS or fully lean into self-managed. 🔹 What Were Evaluating - If self-managed L3 continues to run smoothly under load (network tasks + user tasks + fee paths) - well decommission the RaaS L3 and move forward with self-managed COR L3 as the primary path. 🔹 Why This Is Plausible Now - #Arbitrum #Orbit tooling has matured a lot compared to when we first experimented [--] years ago. - Operational + financial profile"
X Link 2026-01-19T12:28Z [----] followers, [---] engagements
"🛠 DevLog Self-Managed L3: Connectivity Findings & Next Steps Recent stress tests on Testnet-1a (self-managed $COR #L3) surfaced some very practical crossdata center connectivity limits in our setup. 🔹 What Were Seeing - DA (Data Availability) & Sequencer are running in one data center - RPC is running in a different data center - Under sustained load network hiccups between those DCs can: - Disrupt #L3 #L2 data sync - Occasionally force DA resyncs when connectivity between RPC and DA/Sequencer becomes unstable 🔹 What Were Planning - Co-locate core L3 components - Move RPC into the same DC"
X Link 2026-01-21T05:16Z [----] followers, [---] engagements
"Staking APR Adjustment Reminder (Feb 28) Quick reminder that the next scheduled APR step-down for $COR staking goes live on Feb [--] [----] as part of the long-term sustainability plan. 🔹 New APR (from 2/28/2026) - Pool 3: 10% APR - Pool 2: 5% APR - Pool 1: 2% APR Lock periods + all other staking conditions stay the same. This follows the gradual schedule as the network matures and we move into the next phases of Testnet Mainnet. More details: #Cortensor #Staking #COR #DePIN https://docs.cortensor.network/community-and-ecosystem/staking-pool-overview#apr-breakdown 📆 Staking APR Adjustment Next"
X Link 2026-01-23T12:40Z [----] followers, [---] engagements
"🛠 DevLog Testnet-1a & Testnet-1b Status Quick #L3 self-managed check-in. 🔹 Testnet-1a Load & Stability - [--] automated jobs are now running on Testnet-1a (self-managed $COR #L3). - So far things look stable after the move to dedicated servers. - As this holds through the week well ask more node operators to migrate nodes over and gradually increase load. - Live stats: 🔹 Testnet-1b Migration Ongoing - Testnet-1b (internal experimental #L3) is in the middle of its migration to the new dedicated setup. - This environment will stay our breakable lane for DA/Sequencer config experiments and ops"
X Link 2026-01-24T02:28Z [----] followers, [----] engagements
"🛠 DevLog L3 Self-Managed Data Growth (Sequencer Size) Over the last 2448h on the self-managed $COR #L3 weve gotten a clearer picture of how fast sequencer data grows under real traffic. 🔹 What Were Seeing - Current self-managed L3 (Testnet-1a / 1b) sequencer data is already around 20GB after sustained blocks + TXs. - It confirms that a long-running L3 will accumulate significant state as we push more volume and stress tests. 🔹 What Well Explore Next - Track data growth vs TX count / block rate / DA setup more systematically across Testnet-1a and 1b. - Experiment with pruning snapshots and"
X Link 2026-01-25T11:28Z [----] followers, [---] engagements
"🔹 DevLog Data Mobility & Data Module Import/Export (Early Design Thoughts) As we see real data growth on the self-managed #L3 its becoming clearer that storage growth itself isnt the real problem - data mobility is. 🔹 What Were Thinking About - Instead of worrying only about raw disk size were focusing on how to move and reshape state cleanly over time. - The idea is to support import/export for core data modules so we can migrate or hardfork infra without losing the logical network state. Target data modules for import/export design: - Session Data user sessions configs runtime state -"
X Link 2026-01-25T23:11Z [----] followers, [----] engagements
"🛠 DevLog #L3 Data Growth vs Data Mobility (Testnet-1a) Over the last [----] hours on Testnet-1a (self-managed #L3) we finally have some rough numbers on data growth: 🔹 What Were Seeing - Sequencer side is growing around [--] GB/day. - Thats roughly [---] GB/month about [---] TB/year at the current pace. - Most of this is L2 sync / rollup data not pure L3 chain state we fully control. 🔹 Why Storage Isnt the Real Bottleneck - With modern SSD pricing even setups like [--] TB SSD [--] (RAID / multi-disk) per server are very doable. - We can also lean on high-throughput instances (e.g. AWS-type configs)"
X Link 2026-01-26T06:36Z [----] followers, [---] engagements
"🗓 Weekly Focus Phase #2 Stability #L3 Load Expansion & Ranking Refinements This week is focused on keeping Testnet Phase #2 stable while we push heavier load onto the self-managed #L3 track refine validator ranking surfaces for future SLA filters and resume agent-facing endpoint work. 🔹 Testnet Phase #2 Stability & Support - Continue active monitoring and support across all environments (routing miners validators oracles indexers dashboards). 🔹 Testnet-1a Load Expansion + Storage Monitoring - Move more nodes from Testnet-0 Testnet-1a to increase L3 capacity. - Run more manual + automated"
X Link 2026-01-26T07:27Z [----] followers, [----] engagements
"Monthly Node Rewards January [----] (Initial Batch Sent) We've just sent out the first round of $COR rewards to [--] active nodes as part of our monthly node reward program. 🔹 Rewards will be distributed in batches 🔹 More batches rolling out over the next few days 🔹 Recognizing uptime stability & contribution 💬 Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards December [----] (last Batch Sent) Another [--] active nodes"
X Link 2026-01-26T08:55Z [----] followers, [---] engagements
"Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [--] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Initial Batch Sent) We've just sent out the first"
X Link 2026-01-26T22:18Z [----] followers, [---] engagements
"🛠 Devlog Testnet-0 Snapshot & Shifting Load to Self-Managed #L3 Well be taking the next Testnet-0 snapshot for monthly node rewards shortly. Once thats done well start asking node ops to migrate more nodes over to Testnet-1a (self-managed #L3) so we can focus stress tests there. 🔹 Current Testnet-1a State - Explorer: - Stats: 1.8M tx processed 889k blocks 0.5s average block time Daily transactions now in the 350k range 🔹 Next 24h & Node Split Plan - If Testnet-1a stays stable over the next [--] hours well shift from a 50/50 node split (Testnet-0 vs Testnet-1a) to roughly 30% / 70%. - Goal:"
X Link 2026-01-26T23:43Z [----] followers, [---] engagements
"📣 Node Ops Help Us Load-Test Self-Managed #L3 Were shifting more traffic to Testnet-1a (self-managed L3) and would like node operators to rebalance their nodes so that roughly: - 30% of your nodes stay on Testnet-0 (#Arbitrum #Sepolia) - 70% of your nodes run on Testnet-1a (self-managed $COR #L3) This lets us keep Testnet-0 as a stable baseline while really stress testing the self-managed L3 stack (sequencer DA RPC infra) on Testnet-1a. Dashboards: - Testnet-0: - Testnet-1a: #Cortensor #NodeOps #TestnetPhase2 #L3 #DePIN #AIInfra https://dashboard-testnet1a.cortensor.network"
X Link 2026-01-27T00:56Z [----] followers, [---] engagements
"🛠 DevLog /delegate & /validate Iteration with Updated Corgent Endpoints Quick update as we keep iterating on /delegate and /validate on top of Corgent now with the correct router endpoints for both Testnet-1a and Testnet-0 and a clear path into the #ERC8004 ecosystem. 🔹 Testnet-1a Self-Managed L3 (Primary Playground) Internal (Cortensor): - MCP: - REST: External (Bardiel): - MCP: - REST: 🔹 Testnet-0 Baseline Environment - MCP (Cortensor): - MCP (Bardiel): 🔹 What Were Focusing On - Driving more real /delegate flows through Testnet-1a to refine routing retries and tool usage on self-managed"
X Link 2026-01-27T08:41Z [----] followers, [---] engagements
"🛠 DevLog GLM-4.7-Flash Now Live Weve finished wiring in GLM-4.7-Flash and its now ready for wider use across the network. PR: 🔹 What Were Doing - GLM-4.7-Flash support is fully implemented images are built and tested and the PR is ready to use. - A few node operators have already tried it and early feedback looks solid. - Well now start rolling it out more broadly as an additional high-performance model option. 🔹 Why It Matters - GLM-4.7-Flash is extremely strong for agentic coding + tool-calling tasks - in many cases noticeably better than typical OSS baselines. - This gives Cortensor a"
X Link 2026-01-27T10:57Z [----] followers, [---] engagements
"Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [--] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just"
X Link 2026-01-27T20:30Z [----] followers, [---] engagements
"🛠 DevLog Corgent #ERC8004 Agents Wired for Testnet-0 & Testnet-1A Weve finished wiring Corgents external #ERC8004 agents so we can iterate on /delegate and /validate flows from the broader #ERC8004 ecosystem. 🔹 Testnet-0 Corgent Trust Oracle - Sepolia agent: - Base Sepolia agent: 🔹 Testnet-1A Corgent Trust Oracle (Testnet1A) - Sepolia agent: - Base Sepolia agent: These agents are backed by the Corgent #MCP endpoints on Testnet-0 / Testnet-1A and will be used for external /delegate + /validate experiments as we refine Corgent as a shared trust oracle for other agents and apps. #Cortensor"
X Link 2026-01-28T03:16Z [----] followers, [----] engagements
"Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [--] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just"
X Link 2026-01-28T08:34Z [----] followers, [---] engagements
"🛠 DevLog Corgent Primary #ERC8004 Agents + Bardiel #Virtual Layer We now have two layers of #ERC8004 agents wired into Cortensor with Corgent as the primary trust oracle and @BardielTech as the more #Virtual / #ERC8004facing layer. 🔹 Corgent Primary Trust Oracle Agents Testnet-0 Corgent (Trust Oracle) - Base Sepolia: - Sepolia: Testnet-1A Corgent (Trust Oracle Testnet1A) - Base Sepolia: - Sepolia: These are backed by the new router #MCP and #x402 endpoints on Testnet-0 / Testnet-1A and will be the main surface for /delegate + /validate flows as we refine Corgent as a shared trust oracle for"
X Link 2026-01-28T10:24Z [----] followers, [----] engagements
"🛠 DevLog Refined Quantitative / Qualitative Rank Tables (Now on Testnet-0 & Testnet-1A) Weve rolled out the updated validator ranking views to both Testnet-0 and Testnet-1A. 🔹 Whats New - Refined Quantitative Rank Metrics (PoI) and Qualitative Rank Metrics (PoUW) tables with clearer scores + status labels. - Better handling of inactive / never-validated nodes so they dont pollute the healthy set. - Dialog header now shows quick stats (validated / online / inactive active score range) so operators can read network health at a glance. 🔹 Where To See It - Testnet-0 Quantitative: - Testnet-0"
X Link 2026-01-28T12:04Z [----] followers, [---] engagements
"Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just received their $COR rewards - bringing the total to [---] rewarded nodes so far this month. 🔹Rewards will be distributed in batches 🔹More batches rolling out over the next few days 🔹Recognizing uptime stability & contribution Join our Discord to set up a node or get involved: Thanks to all node operators pushing decentralized AI forward - it's $COR or nothing. #Cortensor #AI #DePIN #NodeRewards #COR https://discord.gg/cortensor Monthly Node Rewards January [----] (Batch [--] Sent) Another [--] active nodes have just"
X Link 2026-01-28T22:06Z [----] followers, [---] engagements
"🛠 DevLog Self-Managed #L3 Infra-as-Code (Early Assessment) 🔹 What Weve Proven So Far - Testnet-1A (self-managed L3) is now past 2M tx and 1M blocks running stably under load. - Most of this was brought up and managed via manual setup and config over the last [--] weeks. 🔹 What We Started Today - Walked through the full L3 stack (Nitro DA nodes Blockscout stats proxies monitoring) to: - identify all manual steps - map required configs + secrets and - understand dependencies between components (L2 RPC DA cluster sequencer explorer etc.). 🔹 Where Were Headed - Explore turning self-managed L3"
X Link 2026-01-29T08:49Z [----] followers, [----] engagements
"🛠 DevLog First End-to-End User Task Fee Claims on Testnet-1A 🔹 What We Tested - Started exercising the last mile of the payment flow: claiming user task execution fees from completed sessions. - Ran this on Testnet-1A (self-managed L3) where weve been pushing sustained load tests so node ops actually have real execution fees to claim. 🔹 Early Result - First end-to-end claim of user task execution fees appears to be working as expected on Testnet-1A. - This confirms that the path from: - user task execution validator confirmation node credit claimis now wired through the refactored"
X Link 2026-01-29T12:05Z [----] followers, [----] engagements
"🛠 DevLog Corgent & Bardiel Registered as #ERC8004 Mainnet Agents 🔹 Whats New - #ERC8004 mainnet went live today and weve registered both #Corgent and @BardielTech as mainnet agents so theyre visible in the emerging ecosystem. - For now both are still backed by our Testnet-1A router stack under the hood while we continue to iterate on /delegate and /validate flows during the current testnet phase. 🔹 Mainnet Agent Entries - #Corgent (Primary Trust / Validation Agent) - Ethereum mainnet agent: - #Bardiel (Virtual-Native / Validation & Delegation Surface) - Ethereum mainnet agent: 🔹 Whats"
X Link 2026-01-29T21:09Z [----] followers, [----] engagements
"🛠 DevLog Testnet-1A #L3 Focus Shift (Most Nodes Moving Over) 🔹 Network Status - Weve asked node operators to move all nodes to Testnet-1A keeping only one node on Testnet-0 (L2) as a comparison / fallback. - Testnet-1A is now our primary environment for self-managed L3 benchmarking and monitoring. - Current L3 stats: 1.3M blocks 2.7M tx and 230K daily tx and growing. 🔹 What Were Doing on Testnet-1A - Benchmarking L3 behavior under higher more realistic load (routing validation NodePool utilization). - Tuning reward logic so better-performing nodes are more clearly recognized /"
X Link 2026-01-30T11:09Z [----] followers, [---] engagements
"🛠 DevLog Node Performance / Reward Config Refinement (Now Live on Testnet-1A) 🔹 Weve pushed the updated Node Performance & Reward view to the Testnet-1A dashboard and started using it as the primary way to shape monthly node incentives. 🔹 Live page (Testnet-1A): 🔹 New config knobs - Model Count Threshold minimum number of supported models before a node is considered for rewards. - Min Level (base reward) only nodes at or above a target level earn the base payout. - Must Model Indexes required model indexes that must be online to qualify (already wired into enforcement). - Level Bonus"
X Link 2026-01-31T02:58Z [----] followers, [---] engagements
"🧪 Testnet Phase #2 - 2-Week Check-In (Jan [--] 2026) Its been [--] weeks since Phase #2 kicked off (Jan 17). Theme so far: stable baseline pushing more traffic to self-managed L3 and tightening the trust + routing + incentives loop. ✅ What Happened (So Far) 🔹 Self-Managed #L3 Is Now the Main Test Bed - Testnet-1A (self-managed L3) is at 1M blocks / 2M tx averaging 300K tx/day now running on dedicated servers with DA sequencer and RPC co-located. - Most nodes are being moved from Testnet-0 Testnet-1A targeting roughly 30% (T0) / 70% (T1A) for Phase #2 load. 🔹 L3 Ops Fees & Data Growth Mapped Out"
X Link 2026-01-31T10:52Z [----] followers, [---] engagements
"🗓 Weekly Recap Phase #2 Stability Testnet-1a Load Shift & Ops Improvements This week we completed most planned Phase #2 work with the majority of activity and load now shifted to Testnet-1a (self-managed L3). 🔹 Testnet Phase #2 Stability & Support - Continued active monitoring across routing miners validators oracles indexers and dashboards. - Followed up on abnormal node states with NodeOps; overall Phase #2 stability improved. 🔹 Load Shift to Testnet-1a (Self-Managed L3) - The majority of nodes and testing activity moved from Testnet-0 to Testnet-1a. - Ran additional manual + automated"
X Link 2026-01-31T22:35Z [----] followers, [----] engagements
"🛠 DevLog Testnet Node Capacity Market Conditions & Next L3 Steps 🔹 What Were Doing - Based on community feedback + current market conditions were keeping testnet node capacity intentionally moderate and adjusting it gradually instead of pushing for max scale right now. - Upscaling isnt the bottleneck - spinning up more capacity is straightforward once the timing + market are right. 🔹 What Weve Learned So Far - Last phases (plus current Phase #2) showed the stack can scale: more nodes + more traffic mostly surfaced minor bugs and SLA tuning gaps not hard infra limits. - That means the best"
X Link 2026-02-01T03:49Z [----] followers, [---] engagements
"The quiet lesson from #OpenClaw / #Moltbook-style agent stacks: inference is getting cheap but running agents 24/7 is not. As we spin up fleets of always-on agents Jevons kicks in: lower per-call cost more calls more tools more orchestration and suddenly the real bottlenecks are: - Infra bills (API + hosting + storage) - Trust in what agents did while you werent watching - Long-term memory + state you dont want to recompute every time Cheap inference doesnt end demand it moves value from raw tokens to: ✅ verification / proofs ✅ scheduling & routing ✅ durable memory + context sharing. Thats"
X Link 2026-02-01T08:28Z [----] followers, [----] engagements
"🛠 DevLog CorClaw Agent Economy Experiment (#OpenClaw + #Moltbook + #Clawnch) 🔹 What Were Actually Testing - OpenClaw is the local agent stack were experimenting with. - Moltbook is the agent social network where CorClaw / Cortensor agents can post follow and be discovered: - CorClaw profile: - Cortensor community: - Clawnch is a separate agent-powered launchpad that CorClaw used to deploy a small test token on Base as an on-chain fee stream: - Clanker view: 🔹 Why This Matters - Always-on agents are fun in theory but in practice 24/7 inference is still expensive especially at scale. -"
X Link 2026-02-01T23:36Z [----] followers, [----] engagements
"Everyone loves the agent economy idea. Autonomous agents managing your work Bots trading on your behalf AI companions running 24/7 in the cloud On paper it looks inevitable. In practice it runs http://x.com/i/article/2018140779068395520 http://x.com/i/article/2018140779068395520"
X Link 2026-02-02T02:32Z [----] followers, 37K engagements
"🗓 Weekly Focus Phase #2 Hardening SLA Filters L3 Ops & Agent Surfaces This week continues the Phase #2 push with stability first deeper self-managed L3 iteration and prep work for SLA-driven routing. 🔹 Testnet Phase #2 Stability & Support - Continue active monitoring across routing miners validators oracles indexers dashboards. - Fast follow-ups on abnormal node states and performance regressions. 🔹 Testnet-1a (Self-Managed L3) Iteration + Ops Playbook - Keep iterating Testnet-1a from an operational perspective (reliability throughput failure modes). - Expand and refine the L3 operational"
X Link 2026-02-02T08:36Z [----] followers, [----] engagements
"🛠 DevLog RPC Resilience & Gas Optimization Plans (cortensord) Were shifting some attention over the next couple days to harden how nodes talk to L1/L2/L3 both in terms of RPC usage and gas cost. 🔹 What Were Doing - Reduce overall RPC call volume from cortensord + miner flows and tighten gas usage on the on-chain paths they trigger. - Add support for multiple configurable RPC endpoints so node operators can safely lean on public / free RPC where it makes sense. - Design a simple rotation strategy so operators can supply a list (e.g. up to [--] RPC URLs) and have the node: - fail over cleanly"
X Link 2026-02-02T11:19Z [----] followers, [----] engagements
"🛠 DevLog Thinking Through Multi-RPC Support in cortensord 🔹 Were assessing how to wire multiple RPC endpoints into cortensord / miner without destabilizing the current codepath. - The original design treats the Web3 client as a singleton web3 instance which is now tightly baked into a lot of placesgreat for simplicity not great for RPC rotation. 🔹 Right now were weighing two paths: - A proper refactor where Web3 becomes an injectable dependency (per module / per session) giving us clean multi-RPC support long term. - A more lazy short-term path where we spin up separately named Web3"
X Link 2026-02-02T14:31Z [----] followers, [---] engagements
"🛠 DevLog Cleaner Rank Pipeline for PoI / PoUW Tables 🔹 Rank logic refactor for Quantitative (PoI) & Qualitative (PoUW) views - Simplified the ranking pipeline to use raw validator data directly instead of the older multi-step/derived structure. - Rank rows are now built in two clear stages: - sort by computed rankScore - filter out stale / inactive validators. - Shared static formulas for rankScore and status bands make it much easier to reason about thresholds and tweak them later. - This makes the data model much easier to port into Solidity / NodePoolUtils as an additional SLA signal"
X Link 2026-02-02T23:04Z [----] followers, [----] engagements
"🛠 DevLog Multi-RPC Core Implemented Now in Wiring & Testing Phase 🔹 Status - The core multi-RPC Web3 client (supporting multiple hosts with round-robin rotation) is roughly implemented. - Its not tested end-to-end yet and were in the process of updating the modules that currently rely on the old singleton Web3 instance. 🔹 What This Means - We now have a clear direction: Web3 is effectively becoming an injectable multi-endpoint dependency rather than a single global client. - The remaining work is mostly about wiring and regression: - Swap existing callsites to the new multi-RPC abstraction"
X Link 2026-02-03T07:31Z [----] followers, [---] engagements
"🛠 DevLog Multi-RPC Support: First Attempt & Next Steps 🔹 We tried the first pass of wiring multi-RPC into cortensords core Web3 layer but it didnt stick - once we started plugging it into real modules a few assumptions about single RPC only broke flows. 🔹 The good news: that attempt clarified what we actually need next: - update call-sites module by module - run proper regression tests on each flow - and treat multi-RPC as an incremental rollout not a one-shot refactor. 🔹 Current direction: - keep the new multi-RPC client concept - gradually thread it through the main components (router"
X Link 2026-02-03T12:00Z [----] followers, [---] engagements
"🛠 DevLog Corgent & Bardiel Docs Live (Answering the Whats the Difference Question) A few people have been asking how Corgent and Bardiel differ so weve shipped clearer docs to make the split explicit. 🔹 New / Updated Docs - Corgent overview: - Corgent vs Bardiel comparison: 🔹 TL;DR on the separation - Corgent Infra-native trust & execution oracle for ERC-8004 agents and agent frameworks. Lives closest to the Router + validation stack. Minimal policy-driven /delegate / /validate / arbitration surface. - Bardiel Virtual-first service agent that runs on Cortensor (and often Corgent) but plugs"
X Link 2026-02-03T22:26Z [----] followers, [---] engagements
"🛠 DevLog Rough Sketch for PoI/PoUW-Aware SLA Routing 🔹 New SLA filter path (early sketch) - In parallel with the #RPC and #L3 work weve put together a rough Solidity path for an additional SLA selector that uses PoI/PoUW stats (reserveRandomBySessionSLA2 in NodePool). - If the stats / reputation contracts arent configured it simply falls back to the existing SLA logic so the old behavior is preserved by default. 🔹 Whats different from the old SLA - Previous SLA filters were node-level + spec driven only (CPU/GPU memory model support mostly static capability). - The new path is meant to"
X Link 2026-02-04T03:20Z [----] followers, [---] engagements
"🛠 DevLog Multi-RPC Caching & Next Steps 🔹 Two issues popped up while wiring multi-RPC into cortensord - Multiple RPC instances were reloading the same system config (module addresses version info etc.) over and over. - Every callsite still assumes one RPC endpoint so we need broad replacements to truly support multi-RPC. 🔹 What weve done - Added a simple keystore manager as a singleton cache for module + contract addresses and other shared system config. - On first load one RPC endpoint pulls all runtime addresses stores them in the keystore and subsequent Web3 instances reuse the cached"
X Link 2026-02-04T09:02Z [----] followers, [---] engagements
"🛠 DevLog Integrating Multi-RPC Core Changes into cortensord 🔹 Weve started wiring cortensord into the new multi-RPC Web3 core swapping a small set of calls one by one to make sure behavior stays stable. - The new core supports round-robin rotation across configured endpoints; the rpc_instance: n log line shows which endpoint index is active. - For now were using two test RPC URLs repeated into a 10-endpoint list to verify rotation and basic reliability under real network tasks. 🔹 Early tests look OK but theres still a lot of replacement + regression work ahead. - Most modules still use the"
X Link 2026-02-04T12:41Z [----] followers, [----] engagements
"The Dot-Com Crash the GPU Bubble and Why Distributed Inference Wins the Long Game The last time the world overbuilt critical infrastructure at this scale it wasnt GPUs it was fiber. The dot-com http://x.com/i/article/2019035348077301761 http://x.com/i/article/2019035348077301761"
X Link 2026-02-04T13:32Z [----] followers, [----] engagements
"🛠 DevLog Multi-RPC Integration Progress in cortensord Were mid-way through wiring cortensord into the new multi-RPC Web3 core and testing it in small steps. 🔹 Whats In Place - Web3 now supports round-robin RPC rotation (rpc_instance: n in logs). - A simple config/keystore cache makes sure shared system config (contract addresses versions etc.) is fetched once and reused across all RPC instances. 🔹 What Were Doing - Gradually replacing call sites to use the new multi-RPC helper instead of the old singleton Web3. - After each swap we run light regressions (task metrics errors lag) to catch"
X Link 2026-02-04T23:11Z [----] followers, [---] engagements
"🛠 DevLog Multi-RPC Wiring Pass Mostly Complete 🔹 Most of the web3 call sites in cortensord are now migrated to the new multi-RPC core (round-robin + shared config cache). - Old single-RPC paths are still kept around for safety but the primary codepath is now using the new helper. - Next step: run targeted tests later today (network tasks error rates lag) to make sure behavior matches the previous version under real load. If things look stable well start cleaning up the legacy paths and then roll this into broader testnet usage. #Cortensor #DevLog #RPC #AIInfra #DePIN #NodeOps 🛠 DevLog"
X Link 2026-02-05T05:18Z [----] followers, [---] engagements
"🛠 DevLog Multi-RPC Round-Robin Support for cortensord (PR #211) 🔹 Whats New - Added optional multi-RPC support to cortensord via env flags with simple round-robin rotation across configured endpoints. - New config knobs: - ENABLE_MULTIPLE_RPC_HOST turn multi-RPC mode on/off. - NUM_RPC_HOST how many RPC slots to read. - RPC_HOST_0 . RPC_HOST_N-1 the actual RPC endpoints to rotate through. - When enabled cortensord will cycle through the configured RPC hosts instead of hitting a single endpoint every time. 🔹 Why This Matters - Lets node ops spread load across multiple public / free RPCs"
X Link 2026-02-05T10:08Z [----] followers, [---] engagements
"🛠 DevLog #MCP HTTP Stream Adjustments for Corgent & Bardiel (#ERC8004 Mainnet) 🔹 What We Found - Our current #MCP HTTP stream setup for Corgent and @BardielTech works fine with #OpenAI / #Claude tools discovery. - However some other #MCP clients (including the #ERC8004 scan UI) expect more static advertised agent metadata than we currently expose. - As a result our mainnet #ERC8004 agents are showing up as unhealthy on 8004scan even though the underlying MCP servers are actually functional. - Corgent (mainnet): - Bardiel (mainnet): 🔹 What Well Change - Update the #MCP HTTP stream"
X Link 2026-02-05T11:03Z [----] followers, [----] engagements
"🛠 DevLog Additional SLA Filter (PoI/PoUW) Wiring & Tests Today 🔹 Were spending today iterating on the new additional SLA filter that uses validator-derived PoI/PoUW ranks alongside the existing node-spec SLA. - The first rough implementation is already hooked into NodePool utils; now were focusing on wiring it cleanly through the selection path and making sure it plays nicely with the current SLA levels. - Goal for this round: verify that the new filter can narrow candidates using Quantitative/Qualitative ranks without breaking existing routing or starving valid nodes. - Well run a few"
X Link 2026-02-05T23:05Z [----] followers, [---] engagements
"🛠 DevLog Multi-RPC Rollout Looking Stable (Next: Cleanup & Merge) 🔹 Quick update on the new multi-RPC round-robin path in cortensord (PR #211): so far its behaving stably under testnet load - rotation works and were not seeing obvious regressions on task success or block lag. - If things stay stable over the next few hours well: - Clean up the remaining debug logs and temporary tracing. - Rebuild binaries with the finalized multi-RPC core. - Merge this work into the main branch so it becomes the default code path going forward (still opt-in via env flags). Once merged well ship updated docs"
X Link 2026-02-06T06:08Z [----] followers, [---] engagements
"🛠 DevLog Multi-RPC Support Now Merged into Main Quick follow-up on the multi-RPC work for cortensord: 🔹 Status Update - The new multi-RPC round-robin core is now merged into the main branch. - Behavior is still opt-in via env flags: - ENABLE_MULTIPLE_RPC_HOST - NUM_RPC_HOST - RPC_HOST_0 . RPC_HOST_N-1 🔹 What This Means for Node Ops - Default behavior is unchanged if multi-RPC is disabled. - If enabled cortensord will rotate across multiple RPC endpoints helping: - Spread load across public/free RPCs. - Reduce risk of rate limits or single-provider issues. Next steps: well keep watching"
X Link 2026-02-06T10:17Z [----] followers, [---] engagements
"🛠 DevLog Additional SLA Filter (PoI/PoUW) Moving to Testnet-0 for testing 🔹 Status - The new additional SLA filter is now wired into NodePool utils behind a separate path. - It keeps the existing SLA (spec / capabilitybased) then adds a second pass using Quantitative (PoI) + Qualitative (PoUW) rank data. 🔹 Next Steps - Well deploy this on Testnet-0 and run small controlled traffic slices through the new method. - Main checks: - Routing stays sane vs. the old SLA-only path. - PoI/PoUW filtering narrows candidates without starving good nodes. This is the first step toward SLA that reflects"
X Link 2026-02-06T13:10Z [----] followers, [---] engagements
"Everyone wants the agent economy: Autonomous agents watching markets AI ops-goblins maintaining infra Personal copilots running in the background 24/7 But under all of that hype is a boring brutal http://x.com/i/article/2019376732147707911 http://x.com/i/article/2019376732147707911"
X Link 2026-02-06T22:05Z [----] followers, [----] engagements
"🛠 DevLog PoI/PoUW-Based SLA Filter Now Live on Testnet-0 The new additional SLA filter (using validator PoI/PoUW ranks) is now deployed on Testnet-0 and running in the real routing path for controlled traffic. 🔹 Whats Happening Now - Weve started end-to-end testing of the new SLA method on Testnet-0. - Flow: existing spec-based SLA then PoI/PoUW rank filter as an extra pass. - During this phase well: - Tune thresholds / buckets for the rank bands. - Watch node selection patterns and task success rates. - Confirm it behaves as intended before considering wider rollout. Goal: make SLA reflect"
X Link 2026-02-07T07:59Z [----] followers, [---] engagements
"🛠 DevLog PoI/PoUW SLA Filter Now Active on Testnet-0 & Testnet-1A Weve now rolled out the additional SLA filter (PoI/PoUW-based) to both Testnet-0 and Testnet-1A. 🔹 Current Status - Testnet-0: initial e2e tests are running with the new flow - existing spec-based SLA extra PoI/PoUW rank filter. - Testnet-1A: the same logic is now live and will be our main playground for tuning since it has more capacity and traffic. From here well: - Observe how node selection changes under real load. - Adjust rank bands / thresholds to avoid over-filtering while still rewarding consistently good validators."
X Link 2026-02-07T11:07Z [----] followers, [---] engagements
"🗓 Weekly Recap Phase #2 Hardening L3 Ops Progress & SLA Prep This week we completed nearly all planned focus items for Phase #2 with one item rolling over into next week. 🔹 Testnet Phase #2 Stability & Support - Continued active monitoring and support across routing miners validators oracles indexers and dashboards. - Follow-ups on abnormal node states and performance regressions helped keep Phase #2 stable. 🔹 Testnet-1a (Self-Managed L3) Iteration + Ops Playbook - Further iteration on Testnet-1a from an operational perspective (reliability throughput failure modes). - Expanded and refined"
X Link 2026-02-07T22:35Z [----] followers, [---] engagements
"🛠 DevLog SaltStack Infra for cortensord Fleets Weve just open-sourced our SaltStack-based infrastructure for managing cortensord fleets at scale: 🔹 What It Is - A full SaltStack repo for provisioning configuring and upgrading cortensord across many servers from a single control plane. - Repo: 🔹 Key Features - Centralized node registry to track all your keys configs and roles in one place while still letting you assign specific instances per server. - 3-layer config hierarchy Global Server Instance so you can precisely control everything from chain IDs and RPC URLs to GPU settings and model"
X Link 2026-02-08T07:28Z [----] followers, [---] engagements
"🛠 DevLog L3 Testnet-1A Fee Experiments (Base Fee [--] Stepping) 🔹 This week were starting structured base fee experiments on the self-managed L3 (Testnet-1A) to better understand fee economics vs miner capacity. - Current fee collector (L3 execution fees): - Until now the base fee has been set very low mainly to validate infra + routing under load. - Next step: well increase the base fee by [--] this week and monitor: - Miner/node behavior - Effective weekly fee totals - Relationship between network task volume and L3 fee accumulation - Plan is to repeat the [--] step-up week over week (within"
X Link 2026-02-08T11:29Z [----] followers, [---] engagements
"🛠 DevLog SaltStack Infra: First End-to-End Miner Tests Weve now run end-to-end tests of the new SaltStack setup - not just in theory but all the way through to real miner management: 🔹 What We Just Verified - Brought up a full cortensord miner setup via Salt: install configure systemd service start/stop. - Confirmed the Global Server Instance config hierarchy works as intended for basic flows (chain IDs RPC models GPU flags etc.). - Validated that upgrades + config changes can be pushed centrally without manual SSH. 🔹 Whats Next - This week well start using the SaltStack stack to manage"
X Link 2026-02-08T22:24Z [----] followers, [---] engagements
"https://testnet1a-explorer.cortensor.network/address/0xEa342E0A09FE3CB19245E49422bfea993b6864d1 https://testnet1a-explorer.cortensor.network/address/0xEa342E0A09FE3CB19245E49422bfea993b6864d1"
X Link 2026-02-09T04:06Z [----] followers, [--] engagements
"🛠 DevLog #L3 Base Fee Probe on Testnet-1A Weve kicked off a small economics experiment on the self-managed #L3 (Testnet-1A) to understand what realistic gas flows look like for miners. 🔹 What We Just Changed - Updated the L3 base fee to [----] Gwei on Testnet-1A (applied around Feb [--] 7:55 PM PST). - Fee collector were using as the main gauge: - - Starting snapshot for this experiment: - Collector balance: [----------] $COR at the time of the change. 🔹 What Well Measure Next - Let the network run under normal task load for about a week then: - Calculate balance / [--] days to get a rough daily"
X Link 2026-02-09T04:09Z [----] followers, [---] engagements
"🛠 DevLog #L3 Fee Experiment & Deflationary $COR Mechanics (Testnet-1A) 🔹 Week [--] Base Fee Bump (0.01 Gwei) - L3 base fee on Testnet-1A set to [----] Gwei (Feb [--] 7:55 PM PST). - Fee collector were using as the main gauge: - Balance so far: - Start: [----------] $COR - 5h later: [----------] $COR 🔹 What This Experiment Is For - Use the next [--] days to see how much L3 execution can earn in fees at this base fee level (with [---] nodes active). - Each week well 10x the base fee on testnet then measure again rough curve of: - L3 fee generation - Per-miner economics - How sensitive behavior is to base fee"
X Link 2026-02-09T11:20Z [----] followers, [---] engagements
"🛠 DevLog SaltStack RPC Support for Self-Managed L3 🔹 Weve added RPC setup/management modules to the SaltStack repo so node ops can eventually run their own #L3 RPC endpoints. Salt states: Pillar config: 🔹 Design Notes From last weeks tests an L3 RPC node only needs a relatively small disk since L1/L2 data can come from remote RPCs (e.g. Alchemy). This keeps self-managed L3 more affordable while still giving operators local control over their L3 read/exec path. 🔹 Whats Next Not fully tested in the wild yet - this week well use these RPC modules to stand up an L3 stack via Salt. Goal: once"
X Link 2026-02-09T12:31Z [----] followers, [---] engagements
"🛠 DevLog MCP + ERC-8004 Interop Patch Live on Corgent & Bardiel Weve now applied the MCP + ERC-8004 interop patch to the mainnet Corgent/Bardiel agents and started testing it in-place. 🔹 What Changed - Tweaked MCP stream + metadata behavior so clients that expect more static agent/tool info (like 8004scan) can read health status cleanly. - Redeployed the router nodes behind the mainnet agents with the updated MCP behavior. 🔹 Where Its Live Corgent (mainnet): Bardiel (mainnet): 🔹 What Were Testing Now - Verifying theres no regression with standard MCP clients (OpenAI Claude Tools etc.)"
X Link 2026-02-09T22:53Z [----] followers, [---] engagements
"🛠 DevLog #MCP + #ERC8004 Interop Patch: Testing Regression Monitoring Weve enabled the new #MCP static advertisement payload on the router endpoints behind the mainnet Corgent and Bardiel #ERC8004 agents and are now focused on testing + monitoring rather than adding new behavior. 🔹 Whats Under Test New MCP discovery response that advertises tools categories and protocol info up front so 8004scan can read agent health correctly. Config-gated behavior on the mainnet router nodes backing Corgent and Bardiel (no change to core routing logic or task execution). 🔹 Where Its Running Corgent"
X Link 2026-02-10T02:26Z [----] followers, [---] engagements
"🛠 DevLog #MCP/#ERC8004 Looking Solid Shifting Focus to Data Mobility + /delegate 🔹 MCP / ERC-8004 Status Looks Solid Corgent and Bardiel #MCP endpoints are now showing Healthy on 8004scan and tool discovery works end-to-end. Standard MCP clients (OpenAI / Claude tools) are behaving as expected so were moving MCP to monitor only for now. 🔹 Data Mobility Designing Export/Import for Core Modules Focus this week: admin-level import/export for key data modules so L3s can be migrated or hard-forked cleanly: - Session data - Session queue data - Payment data Node Pool data is largely"
X Link 2026-02-10T10:44Z [----] followers, [---] engagements
"🛠 DevLog Secondary #MCP Endpoints for /delegate Work 🔹 New MCP iteration endpoints (Testnet-1A) Corgent: Bardiel: 🔹 Why Keep the primary MCP endpoints focused on ERC-8004 agents + 8004scan health. Use these secondary endpoints for /delegate experiments breaking changes and fast iteration. 🔹 Whats next Continue expanding the static MCP advertisement (tools + params). Start wiring /delegate as an MCP tool on these endpoints before mirroring anything back to the main surfaces. #Cortensor #DevLog #MCP #Delegate #Corgent #Bardiel #AgenticAI https://router1-t1a.bardiel.app/mcp/messages"
X Link 2026-02-10T22:42Z [----] followers, [---] engagements
"🗓 Weekly Recap Mini Hackathon Ready + Core Improvements As we close out the week Cortensor is now better prepared for real usage - just in time for the first Mini Hackathon. 🔹 Hackathon Infrastructure Ready - Finalized internal hackathon structure & async submission via Discord - Integrated initial tools into: 🔹 Session & SessionQueue Fixes - Resolved key bugs in Oracle v3 affecting task timing & state sync - New Router Node API endpoint /api/v1/create added - enables session creation via RESTful calls - Web3 SDK already supports session creation 🔹 API & SDK Enhancements - Web2 API and"
X Link 2025-06-28T06:30Z [----] followers, [---] engagements
"PSA: There are NO official tokens for Bardiel or Corgent. No #BARDIEL token No #CORGENT token Any token claiming to be Bardiel or Corgent is not launched or endorsed by us treat it as a scam. Always verify via official @Cortensor channels before interacting with anything on-chain. #Cortensor #Bardiel #Corgent #ScamAlert #AgenticAI"
X Link 2026-02-02T23:27Z [----] followers, [----] engagements
"🛠 DevLog #L3 Self-Managed Status Disk Growth & Mock Mainnet Dry Run 🔹 Self-Managed L3 Holding Up (RaaS Paused for Now) - Were decommissioning the RaaS L3 track for the moment the self-managed L3 (Testnet-1A) is stable enough to be our primary focus. - If/when we need it again spinning up a fresh RaaS L3 only takes a few minutes so this is a pause not a permanent removal. 🔹 Real Disk Growth Numbers (Sequencer Focus) - Sequencer disk usage went from [--] GB (Jan 26) to [---] GB (Feb 10). - Thats roughly +100 GB over [--] days [--] GB/day with [---] active miners. - For now the sequencer (plus"
X Link 2026-02-11T04:05Z [----] followers, [---] engagements
"🛠 DevLog Designing Migration Tools for Session / Task Data (Pre-Migration Stage) We havent started any live migrations yet - were focused on designing the tooling that will eventually power #L3 hardforks and module upgrades. 🔹 What Were Working On - Designing off-chain tools that will: - Read Session and Session Queue state via the new admin/paginated getters. - Serialize that state into a portable format for backups / snapshots. - Replay it into a new Session / Session Queue module using the on-chain import functions. - Intended use cases: - Planned hardforks of the #L3 orchestration"
X Link 2026-02-11T22:01Z [----] followers, [---] engagements
"From Cute Agents to a Real Agent Economy: /delegate + /validate 🔹 Two primitives decide whether agent economy is real or just demoware: - /delegate offload work into a scalable execution fabric (fast vs safe tiers) - /validate external checkpoint before actions/payments (VALID / INVALID / RETRY / NEEDS_SPEC) 🔹 On #Cortensor those live in two places: - #Corgent infra-native trust & execution oracle for agents & protocols (#ERC8004 etc.)./delegate + /validate here are minimal policy-driven services sitting closest to the Router + PoI/PoUW stack. - @BardielTech #Virtual-first service agent"
X Link 2026-02-12T02:20Z [----] followers, [---] engagements
"🛠 DevLog Session Data Mobility: First Export Pass We now have the first working export path for Session data on Testnet-0. 🔹 Whats Working - New off-chain Session export tool walks sessions in batches and writes: - one session_id.json per session plus - helper files (user_sessions node_mappings router info manifest etc.). - Counts and basic fields line up with on-chain state (no deep validation yet but structure looks correct). 🔹 Whats Next - Finish the matching import tool to replay these JSON sessions into a fresh Session module (future L3 hardfork / reset target). - If this pattern"
X Link 2026-02-12T04:17Z [----] followers, [---] engagements
"🛠 DevLog Full #L3 Data Mobility Rehearsal (Sessions / Queue / Payments) Weve moved from unit tests to full end-to-end migration / integration testing for the three primary data modules. 🔹 What Were Testing Now - Full export import validation loop on Testnet-0 for: - SessionData (sessions + user mappings) - SessionQueueData (tasks decorations request params task results) - SessionPaymentData (per-node balances pending payments/credits globals) - Flow under test: - Export state from the current contracts into local JSON snapshots. - Run source vs snapshot validation to ensure exports match"
X Link 2026-02-13T20:53Z [----] followers, [---] engagements
"Staking APR Adjustment Reminder (Feb 28) Quick reminder that the next scheduled APR step-down for $COR staking goes live on Feb [--] [----] as part of the long-term sustainability plan. 🔹 New APR (from 2/28/2026) - Pool 3: 10% APR - Pool 2: 5% APR - Pool 1: 2% APR Lock periods + all other staking conditions stay the same. This follows the gradual schedule as the network matures and we move into the next phases of Testnet Mainnet. More details: #Cortensor #Staking #COR #DePIN https://docs.cortensor.network/community-and-ecosystem/staking-pool-overview#apr-breakdown Staking APR Adjustment Reminder"
X Link 2026-01-28T09:43Z [----] followers, [---] engagements
"🛠 DevLog #L3 Ops: Gas Fees Playbooks & Next Steps 🔹 What We Just Did - Ran a base fee / gas fee collection exercise on the self-managed #L3 (Testnet-1A) to make sure accounting + ops flow are sane. - Updated the #L3 execution fee collector to a new address and verified balances before/after the switch: - New fee collector: - Confirmed that #L3 execution fees are accumulating correctly even with a very low base fee right now. - Started drafting internal L3 ops docs around: - changing fee recipients - adjusting base fee / gas params - and doing basic key/collector rotations as part of a"
X Link 2026-01-30T07:46Z [----] followers, [----] engagements
"🛠 DevLog Data Mobility & Import/Export: From Idea Next Steps Last week we outlined data mobility as the real problem to solve on self-managed #L3s not raw disk size. The focus is still on import/export for core data modules (Session Session Queue Session Payment Node Pool) so we can move state cleanly across infra or L3 instances when needed. 🔹 Whats Next (Early Next Week) - Well start mapping concrete import/export shapes for: - Session / Session Queue data - Session Payment balances - Node Pool metadata + reputation - Run small migration drills on testnet/L3 setups to see how cleanly we"
X Link 2026-02-08T02:44Z [----] followers, [---] engagements
"🗓 Weekly Focus Phase #2 Hardening Agent Surfaces & Data Mobility This week continues the Phase #2 push with stability first deeper Testnet-1a iteration and follow-up work on agent-facing endpoints. 🔹 Testnet Phase #2 Stability & Support - Ongoing monitoring across routing miners validators oracles indexers and dashboards. - Fast follow-ups on abnormal node states and performance regressions. 🔹 Testnet-1a (Self-Managed L3) Ops + Load - Completed the full operational checklist for Testnet-1a (baseline stability capacity monitoring). - Move into replay / break-testing mode: intentionally"
X Link 2026-02-09T08:29Z [----] followers, [----] engagements
"🛠 DevLog Corgent & Bardiel #MCP / #ERC8004 Interop Confirmed Healthy 🔹 Whats Now Working 8004scan now shows both mainnet agents as Healthy and fully MCP-compatible. MCP static advertisement payload is live and tool discovery works as expected (10 tools exposed per endpoint). 🔹 Where You Can See It Corgent (mainnet agent): Bardiel (mainnet agent): Corgent MCP: Bardiel MCP: 🔹 What We Verified Standard MCP clients (OpenAI-style Claude tools) still discover and call completions / validation / sessions without regressions. 8004scans endpoint test view can list and ping all tools confirming the"
X Link 2026-02-10T06:40Z [----] followers, [---] engagements
"Staking APR Adjustment Reminder (Feb [--] 2026) Quick reminder that the next scheduled $COR staking APR step-down goes live on Feb [--] [----] as part of the long-term sustainability plan. 🔹 New APR (effective 2/28/2026) - Pool 3: 10% APR - Pool 2: 5% APR - Pool 1: 2% APR Lock periods and staking mechanics stay the same - only the reward rates are adjusting along the previously published curve as the network matures. More details here: #Cortensor #Staking #COR #DePIN https://docs.cortensor.network/community-and-ecosystem/staking-pool-overview#apr-breakdown Staking APR Adjustment Reminder (Feb 28)"
X Link 2026-02-11T06:04Z [----] followers, [---] engagements
"🛠 DevLog Session / Task Data Mobility (First Contract Sketch) Weve started wiring the data mobility layer on the contract side for Session + Session Queue / Task data. 🔹 Whats In Rough Shape - Session Data module - Admin getters with pagination (e.g. getUserSessionsCount getUserSessionsBatch getNodeLastAssignedTimestamp). - First pass of importSession(.) so sessions can be recreated on a new deployment with owner state SLA node lists router metadata etc. - Session Queue / Task module - Migration getters for tasks (getTaskCount getSessionTasksBatch etc.) plus extra getters for decorations /"
X Link 2026-02-11T07:56Z [----] followers, [---] engagements
"🛠 DevLog Session / Task Data Mobility Deployed (Testnet-0 First) Weve now deployed the Session Data and Session Queue / Task modules with their first-pass import/export functions. 🔹 Whats Live - Session module with admin read + importSession(.) support. - Session Queue / Task module with migration-friendly getters and import hooks. Because these imports will reset / rewrite state were treating Testnet-0 (#Arbitrum #Sepolia) as the sandbox for this: - Use Testnet-0 to exercise export serialize import flows end-to-end. - Start building small off-chain tooling to run full migrations (sessions"
X Link 2026-02-11T11:12Z [----] followers, [---] engagements
"💡 Community Spotlight - Monthly Reward: Tools & Infrastructure Each month we reward community members who maintain and improve the critical tools and infrastructure that keep Cortensor's decentralized network running. This month's recipients: 🔹@01_choose Community RPC Node 🔹@ItsMeCryptality Node Monitoring Bot These contributions help node operators and developers onboard monitor and maintain reliable performance - making the network more resilient accessible and efficient. 📌 Access these tools or contribute your own: Thank you to all builders driving the decentralized AI movement"
X Link 2026-02-12T08:13Z [----] followers, [---] engagements
"🛠 DevLog Session Data Mobility Tooling (Export Import Verify) We now have a full export / import / validation loop working for the Session Data module on Arbitrum Sepolia. Target SessionData (new module dry run): 🔹 Whats Working - Export tool walks the current Session Data module in batches and writes local JSON. - Import tool reads that export bundle and recreates sessions + derived mappings in the new Session Data contract. - Validation tool runs twice: - Before import: compares exported JSON vs the original contract state. - After import: compares the new contract vs the original. -"
X Link 2026-02-12T10:38Z [----] followers, [---] engagements
"🛠 DevLog Session Queue Data Mobility & Testnet-0 Reset Weve finished the Session Queue data mobility changes but testing them properly requires a fresh deployment. 🔹 Whats Changing Today The old Session Queue contract on Testnet-0 (Arbitrum Sepolia) doesnt expose the new export/import getters we need. Well redeploy the Session + Session Queue data modules on Testnet-0 so both use the new migration-aware interfaces. As part of this Testnet-0 session / queue data will be reset today. 🔹 What Happens Next Re-run the same flow we used for Session data: Export Session Queue state from the old"
X Link 2026-02-12T20:12Z [----] followers, [---] engagements
"🛠 DevLog Session & Session Queue Migration Modules Live on Testnet-0 Weve deployed the latest Session Data and Session Queue contracts on Testnet-0 (Arbitrum Sepolia) the versions that include full admin import/export hooks so we can run real end-to-end migration tests. 🔹 Whats Now in Place - Both modules expose the new migration functions (export + import) needed to: - Move sessions and queue/task state into local JSON snapshots. - Replay that state into a fresh deployment for hardfork / L3 rotation experiments. - This required redeploying the Session/Queue pair on Testnet-0; earlier"
X Link 2026-02-13T01:48Z [----] followers, [---] engagements
"🛠 DevLog Session Queue Data Mobility: Export Import Validate We ran the same migration pattern we used for SessionData this time on SessionQueueData on Testnet-0. 🔹 What We Just Verified - Exported SessionQueueData (tasks + task results + per-session stats) from the current module into local JSON files. - Ran an export-side validation tool to compare on-chain data vs. the exported files [--] mismatches. - Imported the exported data into the new SessionQueueData module with the admin import functions. - Ran an import-side validation tool against the original source module again [--] mismatches"
X Link 2026-02-13T07:53Z [----] followers, [---] engagements
"🛠 DevLog Sunsetting RaaS #L3 (For Now) Were now comfortable running our self-managed #L3 so were sunsetting the RaaS #L3 by the end of today. 🔹 Why were turning it off - Focus ops and monitoring on a single self-managed #L3 instead of splitting attention. - Reduce extra moving parts while we iterate on fees data mobility and node tooling. 🔹 Not gone forever - Spinning up a new RaaS-based #L3 is now a well-understood path and only takes a short setup cycle. - If/when we need another L3 track (experiments staging or partner work) well bring it back with a fresh deployment. For now the main"
X Link 2026-02-13T09:13Z [----] followers, [---] engagements
"🛠 DevLog Session Payment Data Mobility (Export / Import / Validate) Weve now wired up and dry-run migration tooling for the SessionPaymentData module. 🔹 Whats Working - Built export tooling that pulls session-level payment state (per-node balances pending payments/credits totals globals) into local JSON. - Added import tooling that replays that snapshot into a new SessionPaymentData contract. - Implemented import-side validation: source vs target sessions + nodes now compare cleanly with mismatches = [--] in our test runs. 🔹 Why It Matters - This completes the first migration loop for the"
X Link 2026-02-13T11:08Z [----] followers, [---] engagements
"🛠 DevLog Data Mobility E2E Integration Test (Sessions / Queue / Payments) We just ran a full export import relink rehearsal for the three core data modules on Arbitrum Sepolia (Testnet-0). 🔹 New Target Modules (Post-Migration) - SessionData: - SessionQueueData: - SessionPaymentData: 🔹 Source Modules (Pre-Migration Reference) - SessionData (old): - SessionQueueData (old): - SessionPaymentData (old): 🔹 What We Verified So Far - Exported all session queue/task and payment state from the old modules into local JSON snapshots. - Imported those snapshots into the new contracts above and"
X Link 2026-02-14T02:14Z [----] followers, [---] engagements
"🛠 DevLog Data Mobility: First Pass Complete Weve finished this round of data mobility work by adding the same export/import/validation pattern to NodePoolData and NetworkData on top of the earlier core modules. 🔹 Whats Covered Now - Session data - Session queue / task data - Session payment (pay-per-use) - Node pool data - Network data All five now support: export validate against source import to a new contract validate again with clean round-trips in dry runs. 🔹 What This Enables - Practical #L3 hardforks / rotations with preserved sessions queues payments node registry and network"
X Link 2026-02-14T09:14Z [----] followers, [---] engagements
"🛠 DevLog Prepping Arbitrum Mainnet #L3 Dry-Runs & Experimental (With Data Mobility in Place) Quick snapshot as we get ready for real #L3 experiments on #Arbitrum mainnet. 🔹 Current Cost Snapshot (#Arbitrum #Sepolia) - Approx deploy cost per L3: [------] ETH + [---] $COR - Latest reference deployment tx: 🔹 Why This Matters - #L3 deployment itself is relatively cheap; the main capital cost is validator staking which is fully configurable. - Now that we have basic data mobility across the key modules (Session Session Queue/Task Payment plus NodePool/NetworkData) we can safely: - Exercise the full"
X Link 2026-02-14T10:54Z [----] followers, [---] engagements
"Weve been quietly calling Cortensor AWS for agents for a while now not as a slogan but because the pattern really does rhyme with how AWS changed web apps. Back then AWS basically said: Dont http://x.com/i/article/2022653474115317760 http://x.com/i/article/2022653474115317760"
X Link 2026-02-14T13:37Z [----] followers, [----] engagements
"🗓 Weekly Recap Phase #2 Hardening L3 Ops Progress & Data Mobility Milestone This week we completed most planned Phase #2 work with /delegate and /validate iteration rolling over. 🔹 Testnet Phase #2 Stability & Support - Continued active monitoring across routing miners validators oracles indexers and dashboards. - Followed up on abnormal node states and kept Phase #2 environments stable. 🔹 Testnet-1a (Self-Managed L3) Ops + Replay/Break Testing - Completed replay/break-testing and recovery validation on Testnet-1a. - Ran base fee adjustment exercises and captured learnings into the L3 ops"
X Link 2026-02-14T23:02Z [----] followers, [---] engagements
"🛠 DevLog /delegate & /validate as Core Agent Primitives Since we just shared the longer X article heres the compact builder view recap for Cortensor: 🔹 Two Primitives Were Standardizing - /delegate elastic execution fabric for agents (think AWS for agents): offload heavy work tools and sub-agents to the Cortensor network under clear SLAs. - /validate safety checkpoint before actions/payments: ask the network to judge results (VALID / INVALID / RETRY / NEEDS_SPEC) using PoI/PoUW + validators. 🔹 How This Fits Corgent & Bardiel - Corgent infra-native surface for ERC-8004 and agent frameworks"
X Link 2026-02-15T05:13Z [----] followers, [---] engagements
"🛠 DevLog #Arbitrum #L3 Dry-Run Experiments (Data Mobility Ops & Cost Probing) 🔹 Why Were Doing This - Self-managed #L3 on Testnet-1A has been stable for a few weeks so were confident enough to rehearse the same flows on Arbitrum mainnet. - #L3 deployment costs on #Arbitrum (from #Sepolia tests) are low enough that a few dry-runs are reasonable. - This is not a mainnet launch - just early experiments to build real-world ops + cost data. 🔹 What Well Test - Spin up short-lived #L3s on #Arbitrum mainnet and run through our L3 ops playbook: - Measure gas usage for sequencer batch posting and"
X Link 2026-02-15T09:27Z [----] followers, [--] engagements
"🛠 DevLog #L3 Self-Managed Stress Tests (Testnet-1a 1.04M Blocks / 2.13M Tx) 🔹 More Nodes Moving to Testnet-1a - Weve started shifting additional nodes from Testnet-0 to Testnet-1a (self-managed L3) so we can focus load tests on the self-managed setup. - Target for this week: roughly 30% of nodes on Testnet-0 70% on Testnet-1a now that the monthly reward snapshot is done and distribution is in progress. 🔹 Current L3 Progress (Testnet-1a) - 1.04M blocks produced and 2.13M transactions processed so far. - Recent load sits around 300K daily transactions with block time holding at 0.40.5s. -"
X Link 2026-01-28T00:34Z [----] followers, [----] engagements
"🛠 DevLog Node Performance & Reward Config Refinements (Releasing This Afternoon) Rolling out later today: updates to the Perf / Reward page so incentives track real node performance more tightly. 🔹 Whats New - Refined Node Performance Overview with clearer eligibility counts and average-per-node views. - Model Count Threshold only nodes running enough models are considered for full rewards. - Min Level (base reward) lower-level nodes wont dilute rewards for higher-level better-performing nodes. - Must Model Indexes lets us require specific model coverage (key LLMs) for reward eligibility. -"
X Link 2026-01-30T22:44Z [----] followers, [----] engagements
"🛠 DevLog Moltbook / OpenClaw Experiments & Agent-Spun Cortensor Community 🔹@moltbook + @openclaw Experiments (Inference Providers) - Started light experimentation with #Moltbook + #OpenClaw via the CorClaw profile: - #Moltbook ships with a bunch of inference provider plugins for OpenClaw the closest fit for us right now looks like the OpenRouter API surface. - Over the last few weeks its become clear that a lot of emerging agent frameworks are standardizing on #OpenRouter-style APIs so were planning a deeper look at their spec next month to see how it might map cleanly into Cortensors"
X Link 2026-01-31T06:50Z [----] followers, [---] engagements
"🛠 DevLog Secure LLM Worker Gateway: E2E (Default + DeepSeek Builds) We've expanded Gateway container testing by building images for both the default model and DeepSeek - first step toward rolling this out across all supported models. 🔹 Current Status - Gateway-enabled containers built for default model + DeepSeek. - Local e2e tests show Gateway + LLM worker working side by side. - Cortensord integration confirmed in dev mode. 🔹 Next Steps - Build remaining gateway-enabled images for all models. - Run full e2e trials on remote servers/miners with NodeOps. - Plan stability runs next week"
X Link 2025-09-19T23:20Z [----] followers, [----] engagements
"Cortensord: CPU Usage Control Flag Per request from node operators were adding LLM_OPTION_CPU_THREADS=-1 to optimize CPU usage for AI inference. 🔹 Default: [--] threads [---] retains system default 🔹 Compatible with Docker & subprocess modes 🔹 Now in dev branch (requires a new Docker image) 🔹 Extended testing ongoing this week/weekend This update enhances CPU resource flexibility for improved AI inference performance. Stay tuned #AI #DePIN #Cortensor"
X Link 2025-03-14T03:05Z [----] followers, [----] engagements
"Development Update: Cortensor Desktop App This week we successfully tested the basic MVP functionality across all platforms ensuring Cortensord management works in different environments. ✅ MVP Progress: 🔹 Start/Stop functionality tested on all platforms 🔹 Core protocol management confirmed operational 🛠 Next Steps: 🔸 Wrapping up the remaining basic MVP items next week 🔸 Quick testing with select members from Discord/Telegram 🔸 Evaluating backlog features like auto-installation & additional capabilities 💡 Once the basics are finalized well move toward feature expansion & optimizations."
X Link 2025-03-15T23:21Z [----] followers, 11.8K engagements
"🎉 Phase #2 Testing Milestone Reached Weve officially surpassed [---] active nodes hitting our Phase #2 goal organically 📊 Current Stats: ✅ [---] active nodes ✅ 7880+ sessions processed @aixbt_agent were over [---] nodes The network keeps growing strong. Whats Next Phase #3 testing prep is underwaystay tuned for updates 📌 Track the network live: #Cortensor #AI #DePIN #Phase2Testing #DecentralizedAI #FutureOfAI https://dashboard-devnet3.cortensor.network Phase #2 Testing Network Growth Update The Cortensor network keeps expanding Now at [---] active nodes and growingjust [--] more to hit [---] 🎉 📊"
X Link 2025-03-18T19:21Z [----] followers, [----] engagements
"Development Update: SessionAuth Peer-Level Authentication Were introducing SessionAuth a new smart contract module to handle peer authentication between nodes (e.g. Router Miner) before starting WebSocket communication for real-time LLM token streaming. 🔹 Why SessionAuth Separates peer auth from IAM which handles network-level identity Secures communication without exposing private keys Keeps the system modular and future-ready 🔹 How It Works: A node generates a 6-digit challenge code The peer signs it with their private key The node verifies the signature using the peers public address If"
X Link 2025-03-23T01:41Z [----] followers, [----] engagements
"💡Community Spotlight: Arbitrum Sepolia Faucet Bot for Cortensor Nodes To support smoother Devnet testing weve rewarded @widicrypto with [-----] $COR for building @corfaucetbota Telegram faucet bot tailored for Cortensor node operators The bot provides Arbitrum Sepolia ETH allowing operators to easily cover gas fees and keep their nodes running without friction during testnet phases. 📌 Want to use it 🔗 Full details available on Discord: 🤖 Try it now: Huge thanks to @widicrypto for this practical and thoughtful contribution to the Cortensor ecosystemhelping streamline onboarding and support"
X Link 2025-03-25T22:43Z [----] followers, [----] engagements
"🛠 DevLog: Designing the SessionPayment Module Cortensors Economic Backbone As we expand session capabilities we're introducing the SessionPayment module the financial core for how users pay and nodes get rewarded within Cortensor. 🔹 Why it matters: Handles deposits payments and withdrawals per session Tracks session-specific usage and allocates rewards to nodes Supports scalable secure and transparent payments across the network Enables subscription-style flows for real-time decentralized AI 🔐 With built-in role-based access security guards and auditability it ensures economic trust at"
X Link 2025-04-06T02:24Z [----] followers, [---] engagements
"175M+ $COR Staked 50% of Actual Circulating Supply Weve crossed another major milestone - [------------] $COR is now staked making up 50% of actual circulating supply Massive thanks to the community for your continued belief in Cortensor. This level of commitment is what powers the future of decentralized AI. #Cortensor #AI #DePIN #Staking"
X Link 2025-05-25T06:13Z [----] followers, 13.5K engagements
"Cortensor = The Trust Layer for AI Imagine #OpenAI + #GitHub + #Chainlink - but for decentralized verifiable inference. No central servers. No blind trust. Just output you can verify reward and build on. Thats the difference. #Cortensor #AI #DePIN #TrustlessAI The AI model race will eventually convergewhat truly matters is execution & accessibility. #Cortensor isnt competing in the model wars. Were building the inference layer for all AI. ✅ Scalable ✅ Verifiable ✅ Model-agnostic ✅ Decentralized This is where AI actually runs. The AI model race will eventually convergewhat truly matters is"
X Link 2025-05-31T06:27Z [----] followers, [---] engagements
"🛠 DevLog First Cortensor Telegram Bot via Router API It's live - and a major milestone for decentralized inference. Thanks to quick iteration by Cortensor node operator & community member @beranalpagion we now have a Telegram bot fully powered by Cortensor's Router Node API & miners. ✅ Developer spun up a Router Node ✅ Created a session ✅ Connected the Telegram bot to the REST endpoint ✅ Delivered real-time LLM completions via decentralized inference This marks a true end-to-end flow: from developer Router Node Miner back to the user - without centralized infrastructure. 🔍 What's Next: This"
X Link 2025-06-20T09:54Z [----] followers, [----] engagements
"🛠 DevLog Stat Separation Now Implemented We've implemented the planned stat separation - and testing begins on DevNet5. 🔹 What's Live: - Dual tracking via NodeStats + new SessionStats for session + queue activity - Once validated well fully switch to SessionStats and SessionReputation for user task metrics This allows us to isolate reset and analyze user-level stats without affecting network-level data - a key step toward fairer tracking and future reward scoring. #Cortensor #DevLog #SessionStats #NodeStats #Phase5 #AIInfra #DePIN 🛠 DevLog Stat Separation Planned for Phase #5 As Phase #4"
X Link 2025-06-26T03:58Z [----] followers, [---] engagements
"💡 Community Spotlight - Monthly Reward: Tools & Infrastructure Each month we reward community members who maintain and improve the critical tools and infrastructure that keep Cortensor's decentralized network running. This month's recipients: 🔹@01_choose Community RPC Node 🔹@ItsMeCryptality Node Monitoring Bot These contributions help node operators and developers onboard monitor and maintain reliable performance - making the network more resilient accessible and efficient. 📌 Access these tools or contribute your own: Thank you to all builders driving the decentralized AI movement"
X Link 2026-01-12T10:31Z [----] followers, [----] engagements
Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing