Vibeleaderboard
← All Intel
Intel / post

OpenRouter Adds Hosted Linux Shell Tool for Any Model

Source
x.com
Date
OpenRouter@OpenRouter

1/ Any model on OpenRouter can now run code in a hosted Linux container. Add one tool to a Responses API request. The model writes the script, runs it, and reads back the result. Nothing to install, nothing to host. Meet openrouter:shell. To get started: "tools": [{ "type": "openrouter:shell", "parameters": { "engine": "openrouter" } }]

Context

This thread shows how small the integration is for a developer. OpenRouter says in its post that adding a single tool to a Responses API request lets any model on the platform write a script, run it in a hosted Linux container and read back the result, with nothing to install or host. The example declares a tool of type openrouter:shell with an engine parameter set to openrouter.

That one entry hands the model a working computer instead of asking it to reason about what code would print. OpenRouter's shell tool documentation, read in October, says the commands run in an isolated container rather than on OpenRouter's own infrastructure or the caller's machine, with no outbound internet access by default and a per-command timeout. It bills time at $0.0001 per second, and lists the Responses and Messages APIs as supported and Chat Completions as not.

OpenRouter had already announced the tool in beta six days earlier, alongside a Files API for moving files in and out of the container. That earlier post covers the launch, and this one is the how-to. Only the first post of this thread was collected and its video could not be transcribed, so any detail in later posts is not reflected here.

Terms in this piece · Glossary
  • sandbox — An isolated environment where AI-generated code or agent actions run without being able to touch anything real.
More from OpenRouter
Recommended reads
Comments

Checking sign-in…

Loading comments…