Rate limits

Each API key may make 120 requests a minute. The REST API and the MCP server share the count, so an agent using both has 120 in all.

Past the limit, the API answers 429:

{ "error": "Too many requests. Slow down to 120 a minute." }

Staying under it

• Wait, then retry. After a 429, wait a few seconds before the next request, and longer each time it happens again.
• Spread work out. A loop over 500 conversations should pause between calls rather than send them all at once.
• Ask for less. Filter lists with view, mailbox and q instead of reading every conversation.
• One key per app. Each key has its own limit, so two apps don't use up each other's.

The count is kept close to where your requests arrive, so treat 120 as a ceiling to stay under, not a quota to fill exactly.

Need more for a real use? Write to hello@userookery.com and tell us what you're building.