@dawn-ai/testing

Supported harnesses, fixtures, matchers, deterministic embedders, and runtime helpers for testing Dawn applications.

Use this when: You want to test Dawn agent behavior without making live model calls.

Install

pnpm add -D @dawn-ai/testing vitest

Example

import { createAgentHarness, expectFinalMessage, script } from "@dawn-ai/testing"

await using harness = await createAgentHarness({
  appRoot: process.cwd(),
  route: "/support#agent",
})
const result = await harness.run({
  input: "Say hello",
  fixtures: script().user("Say hello").replies("Hello!"),
})

expectFinalMessage(result).toContain("Hello")

Runtime and stability

@dawn-ai/testing is a supported node-only testing surface. createAgentHarness() temporarily changes process-wide model environment variables and runtime caches; await close() and do not run concurrent harnesses in one process. Fixture replay is deterministic and CI-safe; live recording requires model credentials.

Related

• Testing API reference — exact harness and lifecycle contracts.
• Agent Test Harness — end-to-end agent testing.
• Fixtures and Recording — deterministic replay and optional live recording.
• @dawn-ai/evals — repeatable datasets, scorers, and release gates.

Maturity and support

Dawn is pre-1.0, and its public surface can change. All publishable Dawn packages release together as a fixed group; review the @dawn-ai/testing changelog and upgrading guide before upgrading. For support, use GitHub Discussions; report defects in GitHub Issues.

License

MIT. See the repository license.