FankChen/tracecrate
307 stars · Last commit 2026-09-24
Local-first AI agent trace workbench. Inspect Claude Code, Codex and OTLP logs, compare runs, and export privacy-conscious reports. No backend or API keys.
README preview
# TraceCrate — Your agent ran. Read what happened. **A privacy-first, client-side AI trace workbench. Inspect the sequence, question the patterns, share deliberately.** [](https://github.com/FankChen/tracecrate/actions/workflows/ci.yml) [](https://github.com/FankChen/tracecrate/actions/workflows/pages.yml) [](https://github.com/FankChen/tracecrate/releases/tag/v0.2.0) [简体中文](README.zh-CN.md) · [V0.2 research & design](docs/v0.2-design.md) · [Format guide](docs/formats.md) · [Privacy](docs/privacy.md) · [Contributing](CONTRIBUTING.md)  *Actual V0.2 live-site screenshot from [Pages 34747767675](https://github.com/FankChen/tracecrate/actions/runs/34747767675), visually reviewed; 1440 × 2104 full page. All events and values are synthetic demo data, not benchmark results.* TraceCrate turns files you choose into a searchable timeline, recorded metrics, heuristic diagnostics, and side-by-side comparisons. No backend, no telemetry, no accounts, no API keys. It reads traces; it does not run agents or execute recorded commands. **[Open the live demo](https://fankchen.github.io/tracecrate/)** · **[v0.2.0 release & downloads](https://github.com/FankChen/tracecrate/releases/tag/v0.2.0)** · **[Contribute](CONTRIBUTING.md)** · [Source on GitHub](https://github.com/FankChen/tracecrate). Explore synthetic runs without installing anything, or run locally below. ## Try the story in 30 seconds