mm finds the AI model IDs your code calls and flags the ones that are deprecated or retiring — with the date and the replacement. AI models expire; this catches yours before they do.
curl -fsSL https://llmstatus.ai/install.sh | bash
mm status
Also: brew install randomartifact/tap/modelstatus-cli · npm i -g @modelstatus/cli · try once with npx @modelstatus/cli status
$ mm status
LLM Status — registry live · 16 providers · 466 models
Scanned ~/code/api-service · 6 models tracked
◐ deprecating openai/gpt-4o-2024-05-13 retires 2026-10-23 in 4mo
▲ retiring google/gemini-embedding-001 retires 2026-07-14 in 6d
● ok anthropic/claude-opus-4-7 retires 2027-04-16 in 9mo
mm— full terminal dashboard (scans the repo, browses the registry, live)mm fix --dry-run— rewrites dying model ids to their current replacement, boundary-safe, diff firstmm ci . --fail-on retiring— exits non-zero for CI and pre-commit hooks- GitHub App — fails the check on PRs that introduce a retiring model; one click opens a fix PR
The registry covers 16 providers and updates every 6 hours. Snapshots are Ed25519-signed and verified at install and on every update. Scanning is local: your code never leaves your machine. The CLI sends anonymous usage counts — MM_NO_ANALYTICS=1 turns that off.
The macOS binary is Developer ID signed and Apple-notarized. Linux/Windows binaries carry minisign signatures.
- Radar — what just shipped, what's expiring
- Deprecation tracker — every date we know
- RSS · JSON Feed · retirements only
- Per-model pages: llmstatus.ai/m/anthropic/anthropic/claude-fable-5 (suspended June 12, restored July 1 — the registry tracked both)
Issue tracker and public home for the CLI (@modelstatus/cli) and the docs. The CLI ships as a signed binary; its source is not published here.
Found a bug or want a provider added? Open an issue.
- Site: llmstatus.ai
- Docs: llmstatus.ai/docs · docs.llmstatus.ai
- Changelog: llmstatus.ai/changelog
- Contact: hi@llmstatus.ai