thipdb88

[Cortensor @cortensor]

First Lite.
Then Full.

Q3 2026 → Mainnet Lite
- built on @Arbitrum L2
- more practical and controlled first step
- earlier dedicated-node-backed rollout
- cleaner hosted and product-facing path first

Q4 2026 → Mainnet Full
- built on @Arbitrum Orbit L3
- fuller Cortensor-native path
- broader long-term network direction
- more complete stack beyond Lite

That is the rollout path as we currently see it.

#Cortensor #MainnetLite #Mainnet #Arbitrum

[Cortensor @cortensor]

🛠️ DevLog – Portal Stress Tests Expand Further This Week

A quick follow-up on the current Portal testing path.

🔹 Current direction
- We’re not only increasing token size for the current tests
- We’ll also add the remaining 2 models this week:
- oss-120b
- qwen 3.5 9b

🔹 Why this matters
- That means the Portal path is no longer only being tested with larger prompts / completions
- It also means the multi-model path becomes more complete from an end-to-end testing point of view

🔹 What this helps with
- broader model-routing checks
- heavier token-usage testing
- better end-to-end validation across the hosted path
- more realistic stress-test coverage on Portal, API Gateway, and router-pool behavior

🔹 Current takeaway
- this week’s Portal tests are expanding in 2 directions at once:
- larger token-size / heavier usage
- broader model coverage with oss-120b and qwen 3.5 9b

That should give us a more complete E2E read on the current Portal path.

#Cortensor #DevLog #Portal #StressTest #APIGateway #AIInfra

[Cortensor @cortensor]

🛠️ DevLog – Portal Stress Tests Are Now Using Larger Token Sizes

A quick follow-up on the current Portal stress-test path.

🔹 Current change
- Earlier tests were using fairly minimal prompts and outputs
- We’ve now started increasing both input and output token size so the Portal path is being tested with longer generations too

🔹 Why this matters
- This gives us a more realistic read on how the hosted path behaves once requests are not only frequent, but also heavier in actual token usage
- It also helps us validate the newer token-based usage tracking under more meaningful conditions

🔹 What we’ll test this week
- larger prompt / completion sizes
- more token-heavy request patterns
- different routing / capacity behavior under heavier token load
- other parameter permutations on top of that

🔹 Current takeaway
- Portal stress tests are no longer only about request count
- they are now also pushing more token-heavy usage through the path
- and this week we’ll keep testing those larger-token permutations further

#Cortensor #DevLog #Portal #StressTest #APIGateway #AIInfra

[Cortensor @cortensor]

🛠️ DevLog – Portal V1 Rollout Will Be Gradual: Alpha → Beta → GA

A quick clarification on how we currently think about the Portal V1 rollout path.

🔹 Current state
- Portal V1 is still an early development version
- Most of the current MVP iteration only happened over the last month
- So this should still be thought of as an early product surface, not a finished one

🔹 Alpha phase
- As we move into Mainnet Lite, the initial rollout will likely be an alpha
- That means we will first deploy the unencrypted version
- The reason is simple: it gives us a cleaner way to debug, refine, and improve the product path under real usage conditions

🔹 Beta phase
- Once that alpha path feels stable enough, the next step is beta
- That is where we would turn on private inference / encryption for user input and output
- This would use the router-node privacy feature we developed during Testnet Phase #3

🔹 GA phase
- Then, once the beta path is stable enough, the final step is GA / general availability
- That would mean the product path is mature enough for a broader release

🔹 Why this matters
- The Portal V1 rollout is not meant to be sudden
- It is meant to be gradual, with each phase giving us time to refine the product and reduce risk before the next step

🔹 Simple framing
- Alpha = early Mainnet Lite / Mainnet Full rollout, unencrypted first for debugging and refinement
- Beta = privacy / encryption turned on through router-node private inference
- GA = broader general availability once the beta path is stable enough

#Cortensor #DevLog #Portal #MainnetLite #Privacy #AIInfra

[Cortensor @cortensor]

🗓️ Weekly Focus – Phase #4, Portal Refinement, Mainnet Lite Validation & PyClaw

This week continues the Phase #4 push, with Portal V1 remaining the primary focus while Mainnet Lite validation and PyClaw development continue alongside it.

🔹 Phase #4 – Monitoring, Support & Stats
- Continue monitoring routing, miners, validators, dashboards, indexers, and L3 stats.
- Track stability as Portal V1 and related Phase #4 workstreams continue to mature.

🔹 Portal V1 – Product Refinement
- Continue refining Portal V1 across auth, API keys, usage visibility, request logs, admin surfaces, and overall UX.
- Focus on turning the current MVP into a more polished and operationally usable product.

🔹 Portal V1 – Usage, Analytics & Quotas
- Continue iterating on token-based usage tracking, quota behavior, and analytics surfaces.
- Refine token trends, heatmaps, usage visibility, and quota/reporting accuracy.

🔹 Portal V1 – API Gateway, Router Pools & Scaling
- Continue Gateway → router-pool hardening and operational visibility work.
- Add additional API Gateway instances (already started this weekend) to improve resilience and hosted-path scalability.

🔹 Portal V1 – Multi-Model Routing Tests
- Continue testing multi-model routing across the hosted path.
- Focus on routing behavior, capacity awareness, and model-pool utilization under load.

🔹 Mainnet Lite – Final Baseline Validation
- Mainnet Lite baseline validation is largely complete and looking good so far.
- Re-run Payment Staking E2E once more this week to close out the remaining baseline checks.

🔹 Payment Staking – Regression & Hardening
- Continue validating the recent hardening rollout across Testnet-0 and Testnet-1a.
- Goal is to ensure no regressions before shifting attention elsewhere.

🔹 PyClaw – Dev Path Progress
- Continue PyClaw iteration across workflow, tooling, and repository structure.
- Target is the first public dev/initial commit release by the end of this week, even if still very early-stage.

This week is mainly about Portal V1 refinement and operational hardening, while closing the last Mainnet Lite validation items and pushing PyClaw toward its first public development milestone.

#Cortensor #Testnet #Phase4 #AIInfra #DePIN #Portal #PyClaw #MainnetLite #L3

[Cortensor @cortensor]

🛠️ DevLog – 3 API Gateway Instances Are Now Deployed on Dev0 and Prod

A quick follow-up on the Portal reliability side.

🔹 Current progress
- The 3-instance API Gateway setup is now deployed on both:
- dev0
- prod

🔹 Why this matters
- This gives the hosted path a better baseline for reliability and load distribution
- It also makes the gateway layer less dependent on only a smaller number of instances

🔹 Current takeaway
- The next Portal/API Gateway step is no longer just planning
- the 3-instance setup is now live on both environments and ready for continued monitoring and follow-up testing

#Cortensor #DevLog #Portal #APIGateway #Reliability #API

[HODLHenry @Herbitualz]

@tylerdurdeth @cortensor is the only real decentralized Inference network. Building for 2 years not just tweeting about. Over 50% of $COR circulating supply is staked and waiting for mainnet this qtr

https://t.co/1oXWV9kCfc

[Cortensor @cortensor]

🛠️ DevLog – Portal Follow-Up: More Ops Refinement Now, More Capacity Next Week

A quick follow-up on the current Portal path.

🔹 Current progress
- The Portal admin / ops surface is still being refined as the current tests continue
- A lot of that work is around making the operational view cleaner and more useful as more traffic flows through the system

🔹 What’s also coming next
- Beyond those admin / ops visibility improvements, we also plan to add more capacity next week
- That includes:
- more API Gateway instances
- more router endpoints underneath

🔹 Why this matters
- Better ops visibility helps us understand what is happening
- More gateway and router-endpoint capacity helps the hosted path behave more safely under heavier usage and broader testing

🔹 Current takeaway
- right now = more refinement on admin / ops visibility
- next week = more API Gateway and router-endpoint capacity

So this is both a recap of the current refinement work and a heads-up on the next improvement step.

#Cortensor #DevLog #Portal #APIGateway #Observability #AIInfra

[Cortensor @cortensor]

Cortensor Dashboard is the visibility and operations layer for the network.

It helps users, node operators, and admins inspect what is happening across sessions, tasks, nodes, rewards, config, and runtime state.

That visibility is what makes the network easier to manage and easier to trust.

#Cortensor #Dashboard #AIInfra #DePIN

[Cortensor @cortensor]

🛠️ DevLog – Portal Multi-Model Tests Continue Over the Next 24h

A quick follow-up on the current Portal multi-model testing path.

🔹 Current progress
- The current multi-model setup is in place
- We’ll keep the current test tasks running over the next ~24 hours

🔹 What this helps with
- more routing checks across the current model mix
- more visibility into session usage and route health
- more signal on gateway / pool behavior under continued traffic

🔹 Current note
- We do still need more capacity later for broader testing
- but for now the main focus is to keep this current run going and learn from it

🔹 Current takeaway
- Multi-model routing is in motion
- the next 24h run should give us a better read on how the current setup behaves before the next expansion

#Cortensor #DevLog #Portal #APIGateway #Routing #StressTest

[Cortensor @cortensor]

🛠️ DevLog – Portal Multi-Model Routing: Testing and Refinement Is Next

A quick follow-up on the Portal multi-model path.

🔹 Current progress
- The basic setup for multi-model routing is now there
- That gives us a starting point to test the Portal path with more than one model underneath

🔹 What we’ll focus on next
- more direct testing on the multi-model path
- more routing checks through the Portal and API Gateway
- more UI refinement around how model availability and capacity are shown

🔹 Why this matters
- some of the model-capacity views were still more static/mock before
- now that the model setup is becoming more real, those parts need to become more dynamic too as capacity changes underneath

🔹 Current model focus
- oss-20b
- gemma4
- and some qwen

🔹 Current takeaway
- the base multi-model setup is there
- the next step is testing and refinement
- especially around model routing behavior and making the Portal UI reflect real model-capacity state more accurately

#Cortensor #DevLog #Portal #APIGateway #Routing #AIInfra

[Cortensor @cortensor]

🛠️ DevLog – Portal Multi-Model Routing Works in Basic Form

A quick follow-up on the Portal model-routing side.

🔹 Current progress
- We made adjustments on the API Gateway so different model combinations are easier to reconfigure
- We then tested the latest combo, which currently consists of:
- 2 gemma4
- 1 qwen
- the rest oss-20b

🔹 What we confirmed
- At least in this current setup, model routing is working in basic form
- We were able to confirm routing into gemma4 as well

Reference:
- https://t.co/KVh5Vma5nV

🔹 Current context
- This is still running on only a few nodes for now
- So the current result is more of a basic proof that the routing path works, not the final shape of a broader multi-model pool yet

🔹 What’s next
- We’ll keep testing this further
- Then continue iterating and improving the multi-model routing path from there

#Cortensor #DevLog #Portal #APIGateway #Routing #AIInfra

[Cortensor @cortensor]

🛠️ DevLog – Focus for the Rest of This Week: Portal V1 tests & refinement

A quick recap on what the focus looks like for the rest of this week.

🔹 Mainnet Lite baseline
- The Mainnet Lite baseline and its main critical paths are now working
- That includes the dedicated-node and ephemeral/node-pool baseline checks we wanted to validate first

🔹 What shifts next
- With those baseline checks now in a better place, the focus shifts more toward the Portal side
- The next stretch is mainly about Portal polish, follow-up refinement, and more testing on the hosted path

🔹 Portal-side focus
- more Portal refinement / polish
- more Gateway and router-pool tests
- more stress tests on the hosted path
- more checks on the token-based usage / quota / limit path we deployed earlier this week

🔹 Current takeaway
- Mainnet Lite baseline is no longer the main blocker
- the remaining week is more about tightening the Portal/product path and pushing the newer usage/quota system a bit harder

#Cortensor #DevLog #Portal #MainnetLite #API #StressTest

[Cortensor @cortensor]

🛠️ DevLog – Mainnet Lite Baseline Follow-Up: Ephemeral Node / Node Pool Path Also Worked

A quick follow-up on the remaining Mainnet Lite baseline checks.

🔹 Current result
- The ephemeral-node user-task → node-pool flow also worked
- That means the remaining critical baseline path has now been checked too

🔹 Node pool reference
- Node pool tx path looks clean: https://t.co/p5LyHe7Pjy

🔹 What this means
- At this point, all of the main critical baseline checks for Mainnet Lite are now done and working
- That includes the dedicated-node path and the ephemeral-node / node-pool path

🔹 Current takeaway
- The baseline checklist is now in a much better place
- We’ll continue more prep work after this phase as we move further toward Mainnet Lite

#Cortensor #DevLog #MainnetLite #L2 #Infra #NodePool

[Cortensor @cortensor]

The Dashboard is the real-time operations layer for Cortensor Network.

It gives users, node operators, and admins one place to inspect sessions, tasks, nodes, rewards, config, and runtime state.

That visibility is what makes the network easier to operate and easier to trust.

Mainnet Lite in @Arbitrum is coming in Q3 2026.

#Cortensor #Dashboard #MainnetLite #AIInfra #DePIN

[thipdb88 @thipdb88]

@AlgodTrading $COR Cortensor is the name

[Cortensor @cortensor]

🛠️ DevLog – Mainnet Lite Baseline Checklist: More Ephemeral Nodes Added for the Final Remaining Checks

A quick progress update on the remaining Mainnet Lite baseline checklist.

We’ve now added a total of 5 ephemeral nodes into the node-pool side so we can test the last remaining path more correctly.

This mainly helps with:
- node-pool population
- node-pool lifecycle checks
- the remaining ephemeral-node user-task path

So this is a smaller infra step, but an important one for closing out the last Mainnet Lite baseline checklist items.

#Cortensor #DevLog #MainnetLite #L2 #Infra #NodePool

[Cortensor @cortensor]

🔎 Recap: What is Mainnet Lite vs Mainnet Full?

Mainnet Lite is the more practical and controlled L2 path.

It is taking shape around:
- @Arbitrum L2
- Dedicated-node-heavy serving
- Simpler rollout
- Earlier hosted / demonstration-style path

Mainnet Full is the fuller Cortensor-native path.

It is taking shape around:
- @Arbitrum Orbit L3
- Broader long-term network shape
- Fuller infra / protocol direction
- More complete Cortensor stack

The goal is simple: use Mainnet Lite as the more controlled first step, while Mainnet Full remains the broader long-term network direction.

Mainnet Lite is the earlier rollout path.
Mainnet Full is the fuller Cortensor-native path.

#Cortensor #MainnetLite #Mainnet #Arbitrum

[Cortensor @cortensor]

🛠️ DevLog – Mainnet Lite Follow-Up: Ephemeral Network-Task Path Looks Good, Node Pool Is Next

A quick follow-up on the earlier Mainnet Lite baseline checks.

🔹 Current progress
So far, the ephemeral-node path for network tasks looks good with multiple nodes in the flow.

That means the baseline check around:
- ephemeral node participation
- network-task path
- multi-node behavior

is looking much cleaner now.

🔹 Current reference
Mainnet Lite node pool:
https://t.co/p5LyHe7Pjy

🔹 What still remains
At this point, the main remaining items are:
- node-pool ephemeral-node population
- node-pool lifecycle checks
- user-task flow through ephemeral-node sessions

🔹 Current takeaway
So the remaining Mainnet Lite checklist is becoming more focused now.

The main unfinished area is no longer the basic ephemeral network-task path itself, but the broader node-pool population/lifecycle side and the user-task path on top of that.

#Cortensor #DevLog #MainnetLite #L2 #Infra #NodePool

[Cortensor @cortensor]

🛠️ DevLog – Mainnet Lite Baseline Validation: More Node-Pool Checks Today

A quick follow-up on the current Mainnet Lite baseline checklist.

🔹 Current progress
We’ve now prepared the 2nd and 3rd nodes for the node-pool baseline check on the Mainnet Lite side:
- https://t.co/ldFnTH5i9j
- https://t.co/5fdgkTgnZ0

🔹 What this is for
This is part of the ongoing Mainnet Lite baseline validation around:
- node-pool behavior
- supporting infra/components
- deeper validation of the remaining paths on the checklist

🔹 Main remaining area
The main remaining check is still the deeper testing around the ephemeral-node network-task and user-task path.

🔹 Current direction
So today’s work is mainly about using these additional nodes to keep pushing the remaining Mainnet Lite baseline checks forward and close out more of the checklist.

#Cortensor #DevLog #MainnetLite #L2 #Infra #NodePool