locopilot doctor

Run pre-flight checks.

Source: src/cli/commands/doctor.ts

Synopsis

locopilot doctor

No flags. Always exits with code 1 if any check fails.

Checks

#	Check	What it verifies
1	Ollama reachable	GET http://localhost:11434 responds 200 within 5 s
2	Ollama CLI installed	ollama binary on $PATH; prints version when found
3	Home directory writable	Can write a temp file inside ~/.locopilot/
4	SQLite database exists	~/.locopilot/db.sqlite exists and is non-empty
5	Disk space (> 500 MB free)	fs.statfsSync reports at least 500 MB available — models take several GB each
6	Python 3	python3 --version returns Python 3.x
7	Training adapter	On Apple Silicon: import mlx_lm works. Everywhere else: import unsloth works.
8	Local API server reachable	GET /v1/locopilot/health responds 200 — i.e. locopilot start is running
9	(Pro) Cloud reachable	GET /api/auth/verify with your token returns 200

A failed check still proceeds to the next one — doctor is a status report, not a fail-fast gate.

Sample output

  LocoPilot Doctor

  Platform: darwin / arm64 / Node.js v20.18.0

  ✔ Ollama reachable
  ✔ Ollama CLI installed
  ✔ Home directory writable
  ✔ SQLite database exists
  ✔ Disk space (> 500 MB free)

  ✔ Python 3 (for training adapters)
  ✔ MLX adapter (mlx-lm)

  ✔ Local API server reachable

  Pro: not logged in — run `locopilot login` to enable cloud features

  Free tier ready.

If anything is wrong:

  ✖ Ollama reachable — fetch failed
  ✖ Ollama CLI installed — Not found — brew install --cask ollama
  ...
  3 check(s) failed. Fix the issues above and retry.

Exit codes

Code	Meaning
0	All checks passed
1	One or more checks failed

Use it in CI scripts before running an integration suite that depends on a working stack.