Rate limits - Diagrams.so

Nothing on either plan is metered. The one limit is how fast you call. It is the same on Free and Paid:

Bucket	Limit
Per live key (`dgz_live_`)	60 requests a minute
Per test key (`dgz_test_`)	20 requests a minute
Prompt helpers (`/prompts/enhance`, `/prompts/clarify`), per key	15 requests a minute
Per IP address, across all keys	240 requests a minute

Every response carries RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset (seconds until the window resets). When you go over, you get a 429 with the standard error envelope and a Retry-After header:

{ "error": { "code": "RATE_LIMITED", "message": "Rate limit exceeded. Slow down and retry.", "request_id": "req_..." } }

Handling 429

Back off exponentially and retry. A short pause usually clears it.
Respect the Retry-After header when present.
Spread bursty workloads over time rather than firing in parallel.

Tips

Cache reads

Reads are cacheable. Do not re-fetch unchanged diagrams in a tight loop.

Use idempotency

So a retried create never makes a second diagram.

Limits are generous for normal use. For a high-volume use case, contact support.