Agent Observability
SDKsNode

Vitest reporter

Stream Vitest eval results into the dashboard. Every vitest run becomes one eval_run; every it(...) becomes an eval_case with events, judgments, and failure detail.

The SDK ships a Vitest reporter that turns every vitest run into one eval_run on the dashboard, and every it(...) into an eval_case with its events, judgments, and failure detail.

Configure

Register the reporter and its setup file in vitest.config.ts:

// vitest.config.ts
import { defineConfig } from "vitest/config";
import AgentObservability from "agent-observability-sdk/livekit/vitest";

export default defineConfig({
  test: {
    setupFiles: ["agent-observability-sdk/livekit/vitest/setup"],
    reporters: ["default", new AgentObservability()],
  },
});

Point it at the dashboard and run:

export AGENT_OBSERVABILITY_URL=https://obs.example.com
export AGENT_OBSERVABILITY_AGENT_ID=9c2f7e3d-…
vitest run

Inside a test

Auto-capture and .judge() interception are automatic. The captureRunResult helper is exported for results produced outside the standard session.run(...) path:

import { captureRunResult } from "agent-observability-sdk/livekit/vitest";

it("greets politely", async () => {
  const result = captureRunResult(await session.run({ userInput: "Hello" }));
  result.expect.nextEvent().isMessage({ role: "assistant" });
});

CI metadata (GitHub / GitLab / CircleCI / Buildkite) is auto-detected from standard env vars.

On this page