Rate limits & headers#
Each API key has a per-minute rate limit baked into its tier:
| Tier | Limit |
|---|---|
| slow | 60 RPM |
| medium | 240 RPM |
| fast | 600 RPM |
| embeddings | 1200 RPM |
Limits use a sliding 60-second window, enforced in Redis.
Usage headers#
Every response carries:
| Header | Meaning |
|---|---|
X-RateLimit-Limit |
Your RPM cap. |
X-RateLimit-Remaining |
Requests left in the current window. |
X-RateLimit-Reset |
UNIX timestamp when the window rolls over. |
Retry-After |
Set on 429, in seconds. Honour it — we ratchet enforcement otherwise. |
Need more?#
Email hello@siati.ai — we'll bump it on the spot for non-abusive workloads.
For guaranteed SLAs, see enterprise tier.