OpenDecider - Open decision models that route AI agents without LLM calls
by•
Agent workflows often call an LLM just to decide where a request goes next. OpenDecider answers in one forward pass with a calibrated confidence, sending unsure cases to a fallback. Works with LangGraph, CrewAI, Agno, LlamaIndex, ADK and MCP. Apache-2.0.

Replies
Update: two of the three "coming next" items have shipped.
✅ Guardrails (0.5.0). A prompt guard against jailbreaks and prompt injection, plugged into each framework's own hook. The default guard model, opendecider-small-td, catches as many attacks as Laya's guard with about half the false alarms.
✅ TypeScript client (0.6.0). npm i @opendecider/client. Typed decisions, the router and the guard, with tools for the Vercel AI SDK and Mastra. It has no runtime dependencies and publishes with npm provenance, and each version goes live only after a maintainer approves it with 2FA.
🔜 nano in the browser. We're building this now: opendecider-nano as ONNX, with WebGPU and WebAssembly backends. It already runs in Chrome; we're checking every quantised build against the full benchmarks before it ships.
If you're putting guardrails or routing into a TypeScript or browser app, what would you want from it? Feedback is very welcome.
OpenDecider Focus is now on the Chrome Web Store
It filters your YouTube feed with opendecider-nano running entirely in your browser. You hide the kinds of video you choose, or topics in your own words ("Hide videos about crypto"). There's no server, account or API key, and it works offline after one download.
Install: https://chromewebstore.google.com/detail/opendecider-focus/npjgckljljojppjgikkiiaifoojkoebj
Feedback on what it hides wrongly is the most useful thing you can send me.