MacroDocRefinement

We never store your API keys.

Your API key is saved in your browser's local storage only and is never transmitted to our servers. When using your own model, all API calls go directly from your browser to the AI provider (OpenAI, Anthropic, Google, or xAI). We don't store your keys or your prompts. We do collect anonymous usage and error diagnostics to help keep the app reliable.

Currently using

Default (Free) — Gemini 3.1 Flash Lite (Free)

Free tier (15 requests/minute shared limit)

AI Provider

CH 01 — DEFAULT (FREE)

Default (Free)

Gemini 3.1 Flash Lite via shared proxy

CH 02 — OPENAI

OpenAI

GPT-5.4 series, o3-mini reasoning

CH 03 — ANTHROPIC

Anthropic

Claude 4.5 Haiku, 4.6 Sonnet/Opus

CH 04 — GOOGLE

Google

Gemini 3.1 Flash/Pro (direct API key)

CH 05 — XAI

xAI

Grok 4, 4.1, 4.20 series

Rate Limits

The free tier uses our shared Gemini Flash Lite instance. Signing in with Google lifts your limit:

Signed out15 requests/min
Signed in with Google40 requests/min

Add your own API key for unlimited usage with your preferred model.