Skip to main content
When your agent runs on your own machine, Helix reads its traces from local sources — the transcripts your coding agent already writes, or a JSONL export. Everything stays on disk; nothing is uploaded.

Claude Code transcripts

Reads sessions your Claude Code agent writes under ~/.claude/projects/. Needs no setup: point Helix at the project and it finds the runs.

Codex transcripts

Reads Codex session transcripts (~/.codex/sessions/) in the same way.

Helix sessions

Reads Helix’s own session files under ~/.mutagent/sessions. Helix reads them directly; there is nothing to configure.

JSONL export

Any agent that can emit a trace in JSONL/NDJSON form — one run per line — can be read as a source. Bring your own format; gzipped files are supported.

​
How it connects

Point Helix at the source and it binds it to Evaluate and Diagnose:
.mutagent/config.yaml
A single local source needs no credentials. For a JSONL export, point paths at your files:
.mutagent/config.yaml
See the config reference for every key, including format. With several sources, mark one default: true, or Helix asks which one to use. Run /status in Helix to see whether the source is set up; headless, run helix -p "/status" (it needs a working model key). Each example above is a complete .mutagent/config.yaml; to add a source to an existing file, copy only its sources entry.
Have traces in a hosted observability platform instead? See OpenTelemetry sources.