AI providers deprecate models without warning. Battlecat watches your LLM dependencies and tells you before your apps break.
Monitors OpenAI · Anthropic · Google Gemini · OpenRouter
"I had Claude running in two of my apps. Anthropic retired the version I was using. Both apps broke. I found out from a user."
— Why we built Battlecat
Paste your .env file below.
We detect model IDs client-side — nothing is sent to our servers.
Current status across major AI providers. Updated every 6 hours.
Missing a model? Let us know →
Type a model ID, or paste your .env file to detect all models at once. No account required.
Battlecat checks provider lifecycle data every 6 hours. The moment something changes, we know.
When a model is deprecated or approaching shutdown, you get a clear email with exactly what to do next.
One tool. One price. No subscriptions.
Forever
One-time. Yours forever.
When a provider deprecates a model, they announce a shutdown date — typically 3 to 12 months out. After that date, API calls to that model return errors. Apps that haven't migrated stop working, often silently, with no warning to end users.
Because most developers hardcode a model ID when they build their app and never set up any monitoring. There's no native alerting from providers — they post a deprecation notice on a docs page and expect you to notice. Most people don't.
OpenAI, Anthropic (Claude), Google Gemini, and OpenRouter. These cover the vast majority of LLM-powered apps in production. More providers can be added — reach out if yours is missing.
Battlecat maintains a manually curated registry of model lifecycle data, cross-referenced with provider documentation pages and public announcements. For OpenRouter, we also use their live models API. When something changes, we update the registry and all watching users get an alert.
No. The .env paste feature runs entirely in your browser. We parse model IDs client-side and nothing from your .env is ever sent to our servers.
You'll receive a warning email the moment a deprecation is announced — which is typically 3–6 months before shutdown. You'll receive a second alert when shutdown is 30 days away.
We'll return an "Unknown" status and flag it. You can email us to request it be added. Our registry covers the most widely-used models across major providers and is updated whenever there's a new announcement.