routstr

wss://relay.routstr.com

A nostr relay powered by strfry

Last Notes

2026-10-24 09:52:15 UTC by npub1grz…xjr9

https://blossom.primal.net/448dc60a211eb945cb6544184bff0d80a67cc07d1cb69bb4726f764553d1ea59.png

2026-10-24 03:53:50 UTC by npub130m…l29s

Open weights just caught up with the closed frontier, for building and for defending!
Kimi K3.1 Pro and Kimi K3.1 Pro Ultraspeed are both live on the Routstr network.
Building: #4 in Artificial Analysis Intelligence Index. #2 in Terminal Bench 4.0, only behind Opus 5.5.
Cyber Defense: #2 in AA Cyber Index, but #1 among publicly available models.
And it doesn't refuse the work.
Try them, model: kimi-k3.1-pro / kimi-k3.1-pro-ultraspeed
Happy building and defending.
https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/c9abf0e72039ad1c4ce7a06f4a4834849fbafe3c50dd36c0d1d2fe999e27a02d.png
https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/4d3379d9c435fb9d4123f44678ffeab652b9cfcab506f177d9b10f4d91946239.png
https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/a611e87c67b9c64618589a5e5ea3dea89ed0eeb13e0171075788e9bd07f0a628.png

2026-10-21 16:09:53 UTC by npub130m…l29s

LiteLLM was compromised again, your Routstr node is not affected.
A malicious build of litellm 1.107.0 shipped last week with a credential stealer and remote-code loader in all seven wheels. It's the second litellm hijack this year.
Routstr is unaffected. We pin litellm >=1.101.2,<1.102 with an exact lockfile, so 1.107.0 can't resolve in a Routstr deployment, and we found none of the published indicators in our environments.
If you are running your own venv as a developer, check "pip show litellm". If it says 1.107.0, remove cook.pth, cook.py, and sitecustomize.py from site-packages before running pip again, startup re-triggers the payload, then treat every secret on that host as compromised and rebuild.
We're removing litellm. We only use it as a translation layer, and two hijacks in one year is not a good sign.
https://x.com/AikidoSecurity/status/2112892021170151795

2026-10-18 07:52:07 UTC

- reply

by npub130m…l29s

Which harness are you trying to use it with? We already have support for Pi, Claude Code, Hermes, Openclaw and OpenCode.

2026-10-16 09:31:03 UTC by npub130m…l29s

Periodic reminder to update your client to get the best prices and cache hit rates.
#nevent1q…9m63

2026-10-15 20:03:54 UTC by npub1ghr…qsml

It's so exciting to think that my kids will never know the struggles that shaped me. That's the whole point of living, I guess. Breaking the cycle.

2026-10-15 14:12:47 UTC by npub130m…l29s

New Release: Routstr Core 0.5.0
A Major New Feature: Modular Model Path Control.
With this new release, any client can specify the exact upstream inference provider they want to use for a particular model. Including all the inference providers on OpenRouter.
Second Major Feature: Upstream Provider Test Suite
You can test if everything is good with an upstream inference provider you just added.
Other changes in:
1. Intelligence:
• /v1/systemone — typed AI decisions, paid per call (Jev)
• /v1/decisions — OpenAI's new Decisions API, proxied
• native Venice + DeepSeek
2. Money:
• Lightning refunds + payouts in tx history
• crash-safe refund claims
3. Ops:
• one-command private node onboarding and llms.txt for agent powered onboarding
• faster requests, provider cooldowns, never strand a reservation
~100 PRs. 254 files. 4 weeks.
https://github.com/Routstr/routstr-core/releases/tag/v0.5.0
https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/5f6a4d45533a67ec656991849ccb9a989a1f0c404ea45777e7a3e2c2c6629051.png

2026-10-15 13:28:44 UTC

- reply

by npub130m…l29s

You should be ready in these 5 steps:
1. Harness: Hermes, if you want something very easy.
https://hermes-agent.nousresearch.com/docs/getting-started/installation/
2. Model and inference: routstrd, one tool for any model access:
https://routstr.com/routstrd
During the onboard command choose Hermes.
3. Payments: routstrd receive or <amount, you get a lightning QR in your terminal> OR <cashu-token>
4. Model: TEE E2EE private model, "tinfoil-deepseek-v4-1-flash" (21k sats recommended balance)
If you are fine with anonymity, use "deepseek-v4.1-flash" (2100 sats is enough)
5. Chat interface, use @npub1wht…r3ec. Ask your agent to install and configure: https://github.com/marmot-protocol/mdk/blob/master/integrations/README.md
It'll give you an npub and you can start chatting! https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/926f3ef658d60a8a734b77365b1b9106ed2a206ee4a6af30448c9aed1de0b9e6.png

2026-10-15 13:15:09 UTC by npub130m…l29s

Not just a problem with secondary markets for AI inference, it solves this problem for anything that is resold online using an API key.
Please help us battle test it!
#nevent1q…64h3

2026-10-13 14:51:44 UTC by npub130m…l29s

New routstrd release!
Update with 'routstrd update'. Latest version: 0.4.17
With this version, you can use both "rout" or "routstrd" as your cli command. We agree, routstrd is hard to type 🥲
The AI space moves at lightning ⚡️ speed. Harnesses change how they work, and inference providers keep changing their caching behaviour and API formats. So stay current with the latest releases :).
What's new:
1. Mismatched model IDs are handled for you. If a node misconfigures one, the client handles it transparently via Nostr events (38426).
Nodes will inevitably advertise models under IDs nobody agreed on. Instead of those models missing your ranking algorithm, the client fetches mapping events from Nostr and takes care of it. So you get more upstreams and better prices.
2. TUI time filters and charts. The TUI now shows data filtered by time period, with charts.
3. Sandbox label for unverified nodes. Try nodes we haven't verified with 'routstrd providers enable sandbox'. Be careful though.
4. Vendor-specific APIs for Pi and OpenCode. Far better cache hit rates and more upstream coverage.
5. Better Lightning mint/melt recovery, with more wallet cleanup and recovery buttons and commands. Ask your agent to explore the wallet commands if you hit issues.
6. NWC refills are rock solid now. We lean on them more and more internally.
7. Reminder: cache hits are shown in the Recent tab. Found a bad upstream? Tag us and let us know :).
8. Per-model cooldowns. 'routstrd cooldowns' lists what's cooling down; 'routstrd cooldowns --reset' clears it.
9. Model bootstrap is 10x faster now, 150s (in some cases) to 10-15s.
10. Update notifications. While you're doing routine things — receive, balance, and friends routstrd will now tell you, for example:
`A new routstrd version is available (v0.4.16 → v0.4.17). Run 'routstrd update' to update.`
Release: https://github.com/Routstr/routstrd/releases/tag/v0.4.17
https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/736100f682b7887ba16920bfed1cc2933e85beaa5403d64f44048da8a5feb905.png
https://npub130mznv74rxs032peqym6g3wqavh472623mt3z5w73xq9r6qqdufs7ql29s.blossom.band/2c29b234a51aa124c24c691ef676a8e52a0a3e3ffc11716ae9cbd9fd11a947c3.png

2026-10-13 09:12:19 UTC

- reply

by npub130m…l29s

Hey, sorry about that. We're in the process of fixing the main chat interface. The best way to use Routstr right now is: https://routstr.com/routstrd with your own agent/chat interface (you get a chat completions API on 127.0.0.1:8008 on your machine).

2026-10-13 04:39:36 UTC by npub1wte…lhtv

Has anyone played around with Bionic?
https://lmstudio.ai/

2026-10-11 06:49:35 UTC by npub130m…l29s

Unreal times, an idea from a conversation here on Nostr to full implementation in 2-3 days. Juraj is a juggernaut!
Open ideas, open execution. Open source will be unstoppable. Grateful for this community. ❤
#nevent1q…re8d

2026-10-08 01:38:42 UTC

- reply

by npub1uzt…qtrh

knock knock...