Tuning Engines - Governed AI Runtime

Lets an assistant fine-tune and manage models, run evaluations, and inspect governed agent traces, policies, and approvals through the Tuning Engines platform.

What you can do

The bundled MCP server exposes 60+ tools for training jobs, models, datasets, evaluations, and inference usage, plus runtime governance: traces, policy decisions, approvals, interventions, outcomes, and compliance checks. It can submit fine-tuning jobs on a GitHub repo, estimate costs, monitor GPU usage, and export models to S3. Tenant administration tools manage non-secret registry metadata, team members, and policy dry-runs.

When to use it

Use it when you want an assistant to drive model fine-tuning, evaluation, and inference on the Tuning Engines platform, or to review governance artifacts such as traces, policy decisions, and approval requests. It is not needed for local coding tasks that do not touch that platform.

Requirements

Local Node.js runtime; the package is installed or run via npx. A Tuning Engines account and the TE_API_KEY environment variable are required, obtained by signing up or running the CLI auth login. Network access to the Tuning Engines API is needed, and desktop MCP clients only.