Skip to content

docs: document fair-use rate limits on cloud models #39

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
nchudleigh wants to merge 1 commit into
base: main
Choose a base branch
from
+29 −0
Open

docs: document fair-use rate limits on cloud models #39

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 21 additions & 0 deletions models/language.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,3 +45,24 @@ Ultra-fast inference using Groq's LPU technology. Usage is covered by your licen
| Name | Speed | Benchmark (0-100) | Size | License |
| ---------- | ----- | ----------------- | ----- | ------- |
| Llama 3 8b | 10 | 67 | Cloud | Pro |

## Fair Use & Rate Limits

To keep cloud models fast and available for everyone, Superwhisper applies fair-use rate limits to its hosted cloud models. These are abuse-prevention limits, so normal interactive dictation will not reach them.

Limits apply across all hosted cloud models combined (both language and voice), not per provider:

| Limit | Threshold |
| --------- | ---------------------- |
| Burst | 30 requests per minute |
| Sustained | 300 requests per hour |

<Note>
These limits are subject to change as we tune them. If you have a legitimate high-volume use case, [reach out to support](mailto:support@superwhisper.com) or [bring your own API keys](../security/sensitive-data#use-your-own-api-keys-byok) to use your own provider account.
</Note>

If you exceed a limit, the API returns an `HTTP 429 (Too Many Requests)` response. A `Retry-After` header tells you how many seconds to wait before retrying. Apps built on the Superwhisper API should handle `429` responses gracefully and back off, respecting `Retry-After`.

<Info>
Bring Your Own Key (BYOK) and local models are not subject to these limits, since requests go through your own account or run on your device.
</Info>
8 changes: 8 additions & 0 deletions models/voice.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,3 +51,11 @@ The Nova series of models are a Cloud model hosted by Deepgram.
| Nova 3 | multi | – | 7 | 8 | Cloud | Pro |
| Nova 2 | multi | – | 7 | 7 | Cloud | Pro |
| Nova Medical | en | – | 10 | 7 | Cloud | Pro |

## Fair Use & Rate Limits

Superwhisper's hosted cloud voice models share the same fair-use rate limits as the hosted cloud language models. These are abuse-prevention limits applied across all hosted cloud models combined, so normal interactive dictation will not reach them.

<Note>
See [Fair Use & Rate Limits](../models/language#fair-use-rate-limits) on the Language Models page for the current thresholds and details.
</Note>