Rate limits & quotas
Limits protect the platform and the sites Vector fetches from. They scale with your plan.
Plans
| Plan | Monthly calls | Concurrency | Keys |
|---|---|---|---|
| Free | 1,000 | 2 | 1 |
| Pro | 100,000 | 10 | 10 |
| Scale | 1,000,000 | 50 | Unlimited |
Self-hosting removes these limits entirely — it's your infrastructure.
How it works
- Limits are enforced per API key (and per IP as a backstop).
- Exceeding a limit returns
429 Too Many Requestswith aRetry-Afterheader (seconds). - Monthly quota resets on your billing date; the free quota resets monthly.
Staying under the limit
- Respect
Retry-Afterinstead of hammering. - Cache results you'll reuse — many actors return stable public data.
- Batch where an actor supports multiple inputs.
- Use
vk_test_keys for development so test traffic doesn't burn production quota.