What is Neutrome?

Neutrome lets you expose AI models under your own workspace runtime without building provider routing, tool-loop orchestration, deployment, or runtime-key handling yourself.

You write small TypeScript executors and tools. Neutrome deploys them with your provider configuration and serves standard model APIs. Your clients call one workspace URL with a scoped API key.

What belongs where

• The dashboard is where you create namespaces, models, tools, providers, API keys, deployments, and custom domains.
• User docs teach the workflow and operational decisions.
• The SDK reference describes each public @neutrome/lilsdk and @neutrome/lil-engine export in detail.

The request path

1. A client sends an OpenAI-compatible request to your workspace runtime.
2. The selected model executor receives a normalized program.
3. The executor invokes provider models or other models through the runtime.
4. Attached tools are automatically exposed and executed by the deployed workspace wrapper.
5. The runtime returns the provider-compatible response and records usage.

Continue with the quickstart to create each part once.