NAH AI (@nahruntime) header
NAH AI (@nahruntime) profile picture

Typescript Agent Framework | Coding Agent Join our community: https://nitter.cf/t.co/nr7Q9wBWNw

San Francisco, CA
Joined October 2026
Photo by NAH AI (@nahruntime)Photo by NAH AI (@nahruntime)
we’re working on cloud agents. run our agent in a linux vm with a full filesystem and remotely connect it using your terminal
1
2
102
you can steer a running agent without having to stop the process - nah-ai.com/docs/sdk nah-ai.com/docs/sdk
1
2
8
Replying to @nikolay_dp
Would be cool to support @nahruntime agents 👀👀
1
1
27
Remade the @nahruntime landing page 🫡
2
1
2
94
Potential pricing for @nahruntime cloud agents 👀
2
2
46
New landing page for @nahruntime 👀🚀
1
1
3
83
Working on nah-js which lets you run @nahruntime in the browser with any framework. Currently trying to decouple built in tooling like memory it uses so it’s more configurable
1
1
60
Connected my local @nahruntime agents to the hosted memory it uses instead of running it locally. Now I can share memories from all machines - cogmem.astracollab.app
1
2
4
159
next version of @nahruntime is shipping with cloud sandboxes powered by @boatdotdev 👀🚀
3
1
5
287
Thanks @AniC_dev for all your help👀
next version of @nahruntime is shipping with cloud sandboxes powered by @boatdotdev 👀🚀
1
2
93
One shot landing page for my new project made by @nahruntime using Minimax 3.1 Flash
1
1
2
93
NAH AI retweeted
Replying to @MiniMax_AI
@MiniMax_AI 3.1 Flash is free on NAH! nah.astracollab.app npm i -g nah-ai
Stealth model reveal 🐰 Space Bunny was OpenRouter’s #1 model for 10 straight days, processing 56.7T tokens! The most of any stealth model we’ve run. Space Bunny hopped off OpenRouter on Oct 5. It's back today as @MiniMax_AI's M3.1 Flash, and it's free for the next 72 hours.
2
1
2
199
For some reason sessions with @nahruntime stop at the 1 million token mark so I’m working on extending that
1
2
61
nah’s memory isn’t RAG bolted on. Four tiers (L0 tensions → L1 hot → L2 recall → L3 archive) Recall is deterministic, meaning no model in the loop. The prompt keeps a light index, and full memory text only loads when you name something missing from the transcript.
2
1
2
64