QitOS started with a simple idea: one kernel (the core AgentModule + Engine execution loop), many agents. v0.4 extends that idea to model families. Instead of rewriting examples every time we try a new endpoint, we now resolve:Documentation Index
Fetch the complete documentation index at: https://qitor.mintlify.app/llms.txt
Use this file to discover all available pages before exploring further.
- a family preset (a reusable configuration bundle for a model family)
- a harness policy (the wiring layer that connects a transport, parser, and protocol)
- a transport adapter (the adapter that sends requests to and receives responses from a model API)
- a protocol (the output format the model is asked to follow) and parser (a component that converts raw model output into a typed Decision) pair
gpt-oss, and Gemma 4 without changing the agent implementation itself.