> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.merge.dev/merge-gateway/evals/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.merge.dev/_mcp/server.

# Evals

> Build once. Connect to every API, tool, and LLM. Ship products and agents to production without owning the integration stack.

## Docs

- [Evals](https://docs.merge.dev/merge-gateway/evals/overview.md): Build eval suites from your own test cases, grade them with deterministic assertions and an LLM judge, and use the verdict as evidence for model decisions like migrations.
- [Runs and monitoring](https://docs.merge.dev/merge-gateway/evals/running-evals.md): Eval runs: trials for flaky cases, confidence intervals, paired comparison, alerts, and schedules.
- [Model migrations](https://docs.merge.dev/merge-gateway/evals/model-migrations.md): Prove a candidate model on your eval suite, mirror a share of real traffic to it, then cut over by updating the routing policies you choose.
- [Evals API](https://docs.merge.dev/merge-gateway/evals/evals-api.md): The /v1/evals and /v1/migrations endpoints: gate CI on a suite, sync datasets from your repo, manage schedules, and automate migrations.