AnyModel👻
Ghost Mode
Zero-retention API keys. Your prompts and responses are never written to disk — only a token counter remains, for billing. Same price, full privacy on our side.
🚫Nothing stored
With a Ghost key, request and response bodies are never persisted. We keep only the token count needed to bill you — never your text.
🔑Per-key, opt-in
Flip Ghost Mode on any key. Mix a Ghost key for sensitive work with a normal key for the rest — same endpoint, same models.
🧾Verifiable, not vibes
Don't trust us blindly — verify. Ghost responses carry a signed no-store receipt, and you can ask our API exactly what we hold on your key.
How it works
- 1. Create an API key and turn on Ghost Mode.
- 2. Use it anywhere — direct API, Codex, Cursor, any OpenAI-compatible client.
- 3. Every request with that key skips body storage on our side. Only the token tally is recorded.
An honest note. Your prompt is still sent to the model provider (OpenAI, Anthropic, Google) to generate a response — that's how any LLM works. Ghost Mode is about what we keep: nothing but the token count. We don't claim "100% privacy" — we claim zero retention on our side, and we let you verify it.