Limits
Default project quotas and API rate limits, and how to raise them.
Limits protect capacity for everyone and stop runaway automation from creating more than you meant to. Defaults are below; raise any of them from the console or by contacting us.
Project quotas
| Resource | Default |
|---|---|
| On-demand GPUs per project | 16 |
| Nodes per project | 8 |
| Block storage per project | 100 TB |
| Access tokens per project | 50 |
Reserved clusters are not limited by on-demand quotas; their size is agreed when the reservation is made.
API rate limits
| Scope | Limit |
|---|---|
| Read requests | 600 per minute per token |
| Write requests | 60 per minute per token |
| Attestation fetches | 120 per minute per node |
When a limit is exceeded the API returns 429 with a Retry-After header in seconds.
Quotas in this documentation are proposed defaults for the private preview and may change before general availability.