Rate Limits Reference
Command cooldowns and HTTP API rate limits.
Rate limits reference
Command cooldowns
Cooldowns are per user, per command. They only apply to commands that set one; most information commands and check-ins have no cooldown beyond Discord's own interaction limits.
| Command(s) | Cooldown |
|---|---|
/coping * | ~8 seconds |
/ghostletter | ~8 seconds |
/joke, /compliment, /trivia, /wouldyourather | ~5 seconds |
/guildsettings | ~5 seconds |
/guilddebug, /guildcontext | ~10 seconds |
/version, /stats | ~10 seconds |
/ping, /about, other info | ~5 seconds |
| Owner commands | ~3 seconds |
Exceeding a cooldown returns a short "try again in N seconds" reply. There is no server-wide command quota.
Crisis screening
Crisis screening runs at most once per triggering message (a /crisis command, the context
menu, a DM, or a mention). It is not run continuously.
HTTP API
| Scope | Limit |
|---|---|
| All endpoints | 60 requests / minute / IP |
POST /v1/chat | 20 requests / minute |
POST /v1/feedback | 5 requests / minute |
Requests over the limit receive 429 Too Many Requests.
Anthropic
Upstream Anthropic rate and token limits apply to AI features. When Anthropic is unavailable or rate-limited, Mellow uses fixed fallback responses rather than failing.