Use the API · 5 of 5
Errors and limits
What each error means, which to retry, and how many requests can run at once.
- Who can do this
- Everyone
- Reading time
- 1 min read
Errors come as JSON with a detail holding a type, a code, a message and the request_id.
| Status | Code | What to do |
|---|---|---|
| 401 | missing_api_key, invalid_api_key | Check the key and the header. |
| 403 | insufficient_permissions, forbidden | Give the key the endpoint's permission, or allow the address. |
| 404 | voice_not_found, model_not_found | Check the ID. |
| 400 | validation_error codes |
The param field names what's wrong.
|
| 402 | insufficient_credits | Buy credits, change plan, or raise the key's monthly cap. |
| 429 | concurrent_limit_exceeded, system_busy | Retry after a moment. Nothing was charged. |
| 500, 503 | internal_error, service_unavailable | Retry. Nothing was charged. |
How many at once
Your plan sets how many requests can be speaking at the same time. More than that are refused with 429 straight away, before anything is held, so they're safe to retry; the SDKs do it for you. Test keys and repeats with a seed never count.
Still stuck?
A person at Navoras reads every message and replies by email. Name your workspace, and the request ID if it's about one.