Which of your LLM models are about to stop working?
Paste a public GitHub repo or some code. We flag model IDs that are retired or retiring on OpenAI, Anthropic, Gemini, Bedrock and Vertex, and parameters the new models reject, with the official date and the replacement.
Dated, sourced, current
Every row links to the provider's own deprecation page and says when we last checked it. Bedrock and Vertex keep their own calendars; we track those too.
Catches the quiet breakers
Opus 4.7 and later return HTTP 400 on a non-default temperature, top_p or top_k. We flag it next to the model, and when the recommended replacement would reject it.
Fails CI before prod does
A free GitHub Action and a one-file CLI fail the build when a model retires within N days. Annotations point at the exact line.
Free vs paid
| Repo scan, CLI, GitHub Action, shutdown calendar | Free |
| Migration report: old vs new model on your saved prompts, side-by-side outputs, a patch | $199 once |
| Auto-PR: a pull request with the diff report every time one of your models gets a shutdown date | $49 per repo / month |