Getting 429 errors in SillyTavern or JanitorAI? What to do
Updated
A 429 means "too many requests". On UnsavedAI it comes in three kinds: too many of your own messages waiting, the shared queue being full, or more than your per-minute limit. Most come from swipes, several tabs or auto-retry firing extra requests while a reply is still being written. Let replies finish, wait half a minute, then send one message.
Try it now, free, no account neededStart a story with The NarratorAny story you want, you chooseHow the queue works
Free writes one reply at a time and Priority two. Anything extra you send waits its turn, and up to four more of your requests can wait. Requests from everyone else wait in a shared line, with Priority ahead of Free. A 429 means one of those limits was reached; nothing is wrong with your key or settings.
The three messages and what each means
- "Too many requests from your account are waiting. Try again shortly." You already have a reply being written and four more waiting. Stop or finish them before sending more.
- "The request queue is full. Try again shortly." Everyone is busy. Wait about thirty seconds and send again.
- "Your account requests-per-minute limit is currently reached." Something sent more than 60 requests in a minute, or 300 on Priority. It clears within a minute.
A related one, "Your earlier requests were still running and this one waited too long. Try again shortly.", means a request queued behind your own work for too long. The fix is the same: fewer things at once.
What usually causes it in roleplay apps
- Swiping several times while a reply is still generating. Each swipe is a new request.
- The same chat open in two or three tabs, or JanitorAI and SillyTavern running at once.
- Auto-retry or "regenerate on error" settings that resend immediately, turning one failure into several.
- Extensions that make their own requests, such as automatic summaries or image prompts, firing alongside your message.
- A shared server, such as an Open WebUI install, with several people on one account.
Making them stop
Wait for a reply to finish before swiping. Close duplicate tabs. Turn off auto-retry, or make it wait several seconds between tries; instantly resending in parallel only makes the wait longer. If you use background extensions, let them run after a reply rather than with it.
If you need two chats going at once regularly, that is what Priority's second slot is for. It also puts you ahead of Free in the shared queue, though nobody can promise instant replies when everything is busy. Our streaming and retries guide covers retry timing for developers.
Questions people ask
Do extra API keys get around a 429?
No. Limits belong to the account, so every key shares the same slots and per-minute allowance.
Is there a daily limit?
No. Neither plan has a daily token cap; the limits are on how many requests run or wait at once and per minute.
I sent one message and still got a 429. Why?
Usually something else on your account was still running: another tab, a swipe or an extension. If not, the shared queue was full; wait half a minute and try again.
