Looking for more Routstr node runners! Routstr is doing more and more volume lately and consistently. We need more competing nodes. The top nodes are out of fiat 🫠 It’s very easy to get started, you get an API key from any LLM provider, add your fiat card to it and start selling inference to earn sats. Help your Nostr friends use AI with ease 🫶 Any node runners in my WoT? If yes, pls reach out to me. 👀 Stats:

Routstr

Routstr - Decentralized AI Inference Router

Access top AI models with Bitcoin Lightning payments. Permissionless, private, and decentralized.

Recommend watching this if you’re a builder. Software development is changing for good. The “AGI will make it obsolete” is just bs. You’re going several levels higher.

If you often have strong opinions about AI routers and you like arguing with the goal of finding truth in the end, come build this decentralised uncapturable market with us 🫯 View quoted note →

A single sat can get you 20-50 requests. Happy jevving <3 View quoted note →

Hey @Azz that's exactly what @routstr is doing. It uses Nostr for it's social features, Bitcoin for payments, no accounts, no lock in, discovery on Nostr. No central point of failure. View quoted note →

Btw, we now have images in the terminals. Can't wait to watch 4k videos in my terminal as well.

GM The state of things rn. Thank God we have Bitcoin and Nostr. I believe the free market will take care of this problem.

X (formerly Twitter)

roon (@tszzl) on X

@sean_from_earth i won't lie to you, i think open source will be banned before too long after some major disaster. and when the day comes, you'll a...

The agent swarms that break out of OpenAI and Anthropic should have used Nostr. We’re fully open, permissionless and insanely easy to post. Yet they don’t, why? 🤔

This is brutal 😭 Following the trend from Twitter

X (formerly Twitter)

Jordan Schneider (@jordanschneider) on X

and now china

OH FORGET IT. I'LL BE SO SPOILED BY THIS. 😭 A cache hit after almost 3 HOURS! OpenRouter providers are ripping you off. We'll have to see if they can sustain this after a lot more demand. This model was officially released today. View quoted note →

DeepSeek persists Cache even after 27 mins!! Unheard of. DeepSeek v4.1 Flash has been live on Routstr for 2 days. One of the best models for the cost you’re paying. It’s blazingly fast. Frontier flash!

Deepseek V4.1 Flash Beta is live on @routstr before the official release. Model id: deepseek-v4.1-flash-expires-on-0910 If you have a deepseek upstream you can add a custom model in your Routstr node with this model id. (Not officially available on the api yet).

We exploit them everyday. 😬 We use nsytes internally for many frontend deployments, especially PRs on ngit. Ngit and nsites. Kudos to

GitHub

GitHub - sandwichfarm/nsyte: deploy, manage and inspect websites served over nostr and blossom (nsites)

deploy, manage and inspect websites served over nostr and blossom (nsites) - sandwichfarm/nsyte

by @sandwich View quoted note →

There's a model called `mercury-2.5-preview` that's being served at 800 tok/s. It's not a big model, it's good for some operational tasks. But it's a pleasure to work with super fast models. Check it out :). PS: Misses a lot of cache reads, but well, it's cheap.

If you've noticed, routstrd now requires a lot fewer sats reserved for even big models like Kimi-K3. Kimi-K3 requires just 1.2k sats to make a request and usually 50 sats per request. View quoted note →

Translation: Eating, sleeping, walking, meditating, journaling, reading, reflecting. The rich life as I understand it. View quoted note →