Back to directory

Latitude

An open-source observability platform for AI agents that groups failing traces into tracked signals, dispatches a coding agent with the relevant context to fix the underlying bug, and replays fixes against real failures.

Latitude

Latitude observes AI agents through one line of telemetry that captures every trace, including multi-turn sessions, tool calls and full execution paths, ingested over OpenTelemetry. Failing traces are automatically grouped into tracked signals showing status, size and trend, so a team can see what is breaking and how often.

When a signal is worth fixing, Latitude dispatches a coding agent such as Claude Code or Cursor with the full context, sample traces and a deep link, so it can write a minimal fix and open a pull request; the fix is then replayed against the real failing traces to confirm it holds. The same functionality available in its web interface can also be reached from a coding agent through an MCP server and a CLI, and telemetry works with most model providers and frameworks, including OpenAI, Anthropic, Bedrock, the Vercel AI SDK and LangChain, plus any OpenTelemetry-compatible application.

More in AI evals

See all 269
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • JevBench

    benchmarkheaven.com

    A reproducible benchmark from Benchmark Heaven that ranks Jev-class typed decision models on intelligence, calibration, speed, and cost, with a published harness, scoring rules, and results.

  • A co-evolutionary adversarial red-teaming tool for AI agents that uses quality-diversity evolutionary search to autonomously breed edge-case scenarios, testing whether an agent with tool access can be made to hallucinate, deadlock or leak funds.