# agentminutes

> Parse native agent harness session logs (Antigravity CLI, Claude Code, Codex CLI) into one unified, comparable event schema.

## Documentation

- [Home](https://agentminutes.dev/index.md): Parse native agent harness session logs (Antigravity CLI, Claude Code, Codex CLI) into one unified, comparable event schema.
- [Documentation](https://agentminutes.dev/docs/index.md): Guides and reference for agentminutes.
- [CLI](https://agentminutes.dev/docs/cli/index.md): The convert, sessions, detect, stats, and drift commands.
- [Design Notes](https://agentminutes.dev/docs/design/index.md): Loud failure, post-hoc parsing, streaming, and the explicit registry.
- [Session Discovery](https://agentminutes.dev/docs/discovery/index.md): Finding transcripts by cwd, session ID, and time window.
- [Example: Comparing Two Harnesses](https://agentminutes.dev/docs/example-comparison/index.md): A worked example, from capture to a side-by-side of tool use, context, and cost.
- [Harness Support](https://agentminutes.dev/docs/harnesses/index.md): Supported harnesses, native formats, and validation coverage.
- [Installation](https://agentminutes.dev/docs/installation/index.md): Homebrew, npm, PyPI, go install, and prebuilt binaries.
- [Go Library](https://agentminutes.dev/docs/library/index.md): Whole-file parsing, streaming, transforms, and the ACP projection.
- [Quickstart](https://agentminutes.dev/docs/quickstart/index.md): Install agentminutes and parse your first transcript.
- [The Schema](https://agentminutes.dev/docs/schema/index.md): The normalized event schema and the guarantees behind it.
- [Subagents](https://agentminutes.dev/docs/subagents/index.md): How subagent transcripts are stored, how they tie back to the parent session, and how to analyze a whole task.
- [Comparing Token Counts](https://agentminutes.dev/docs/token-comparison/index.md): What token numbers mean per harness, which ones are comparable, and which comparisons are traps.
- [Use Cases](https://agentminutes.dev/docs/use-cases/index.md): Why parse and compare agent transcripts across harnesses?