add-atomic-chat-tool

You can expose local Atomic Chat models to a container agent by deploying an MCP server with an OpenAI-compatible API, which registers tools, forwards host environment overrides, and enables on-host inference.

Yes, an MCP server connects automated container agent tasks to local models by copying the MCP source, updating tool allowlists, and forwarding host environment overrides for private, offline AI execution.

Before integrating local Atomic Chat models, you must install Atomic Chat on the host and enable the Local API Server on port 1337 so the container agent can successfully list models and generate responses.

Yes, the Atomic Chat MCP server integration supports OpenAI-compatible APIs, allowing development workflows to leverage cost-efficient, private AI execution inside containerized environments.

Using local models with a container agent provides on-host inference, ensuring privacy, offline capability, and cost-efficient AI execution suitable for development, testing, and quick model probing.