> ## Documentation Index
> Fetch the complete documentation index at: https://docs.audivo.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Limits

> What each plan allows: requests per minute, jobs in flight, selection sizes, and the audio caps everyone shares.

Every limit below is enforced in the API or at its edge, and each one has a
typed error when it bites. Prices and each plan's included credits are on
[Credits](/credits).

## Per plan

| plan    | requests per minute | open jobs at once |
| ------- | ------------------: | ----------------: |
| free    |                  10 |                 3 |
| hobby   |                  60 |                10 |
| startup |                 150 |                30 |
| growth  |                 400 |                60 |
| scale   |               1,000 |               100 |

| plan    | episodes per selection | chart size |
| ------- | ---------------------: | ---------: |
| free    |                      3 |         10 |
| hobby   |                     10 |         20 |
| startup |                     30 |         30 |
| growth  |                     60 |         40 |
| scale   |                    100 |         50 |

* **Requests per minute** is enforced at the edge with a token bucket. Over
  it, the answer is `429 rate_limited` with a `Retry-After` header; the same
  request succeeds once the bucket refills. `X-RateLimit-Limit` on every
  response states your plan's number.
* **Open jobs** are jobs reserved and not yet terminal, counted across all of
  your groups. A confirm that would exceed the cap is refused as
  `429 concurrency_limited` naming the cap and the room left, before anything
  is reserved. Cached reads do not count.
* **Episodes per selection** bounds one quote: a selection larger than the
  plan allows is clamped and the clamp is reported on the quote, never
  silently dropped. No plan limits how far back into a show's catalogue you
  may reach; name the episodes you want.
* **Chart size** bounds how many shows a category chart contributes.

## Shared by every plan

| cap                        | value                                   |
| -------------------------- | --------------------------------------- |
| Audio duration per episode | 10 hours                                |
| Audio file size            | 5 GB                                    |
| Transcript served inline   | 5 MB per rendering, then a 24-hour link |
| Idempotency key memory     | 24 hours                                |
| Quote validity             | 15 minutes                              |
| Live API keys per account  | 25                                      |

An episode over the duration or size cap is refused as `duration_exceeded`
or `size_exceeded` at validation, with nothing reserved.

## Reading your own numbers

An API-key caller reads its balance and reservations from every quote
response (`balance_credits`, `reserved_credits`). The usage and limits pages
are part of the dashboard and answer only a signed-in session, never an API
key.