Limits and Quotas
Dregs caps how much each plan can ingest, how fast a client can call the API, and how large a request can be. This page lists those caps and what happens when you reach one.
A new account starts on the free trial. Starter, Growth, and Advanced are the paid plans. Prices, and a shorter list of the plan allowances, are on the pricing page.
The caps
| Limit | Value | What happens |
|---|---|---|
| Included active identities | Starter: 100, then $0.06 each. Growth: 5,000, then $0.05 each. Advanced: 10,000, then $0.04 each. | An active identity is one that was tracked and scored in the month. Passing the included count does not block new identities. The extra identities are billed at the plan's per-identity rate at the end of the billing cycle. |
| Events per month | Free trial: 10,000. Starter: 1,000,000. Growth: 5,000,000. Advanced: 10,000,000. |
At 80% of the allowance, the billing address gets one warning email for that period.
Once the count reaches the allowance, the next new event is refused with HTTP 402 and
status quota_exceeded. The dashboard says event syncing is paused, and the
billing address is emailed that the event limit has been reached. Syncing resumes next
period, or when you upgrade. Sending an event id that was already stored does not count again.
|
| Checks per month | Starter: 500,000. Growth: 2,500,000. Advanced: 5,000,000. | A check is a score lookup for an identity you already track, through the API, the dashboard, or MCP. It reads scores Dregs has already computed, and another lookup is not another charge. Authenticity Check API calls are separate and count against the operations quota. |
| Operations per month | 10,000 per billing period, unless the plan sets a different cap. An account with no active subscription gets 10,000 per calendar month. |
A signed-in POST /api/tools/authenticity-check (and the older
/api/tools/identity-check path) past the cap returns HTTP 402 with status
quota_exceeded and the message "Monthly operations quota exceeded for your plan."
A signed-out call does not spend this quota.
|
| Team members | Free trial and Starter: 3. Growth: 5. Advanced: 10. | Inviting another member past the cap is refused with HTTP 409. The count is enabled members on the team. A trainer on the account is not counted. |
| Badge rules | Free trial and Starter: 10. Growth: 20. Advanced: 30. | Creating another rule past the cap is refused with HTTP 409. |
| Escalation rules | Free trial and Starter: 10. Growth: 20. Advanced: 30. | Creating another rule past the cap is refused with HTTP 409. |
| Notification channels | Free trial: 3. Starter: 1. Growth: 3. Advanced: 5. | Creating another enabled channel past the cap is refused with HTTP 409. Disabled channels are not counted. |
| Custom analyzers and custom datasets | Free trial: 0. Advanced is the plan that includes them. Starter and Growth do not list them. | Creating one past the plan cap is refused with HTTP 409. |
| Event history | Free trial: 30 days. Starter: 90 days. Growth: 180 days. Advanced: 365 days. | Your plan's history window sets how far back you can see and query events. Older events aren't shown in the dashboard or returned by the API, and aren't used in scoring. |
| Free trial and grace period | 14-day trial, then 14 days of grace. No credit card is required to start. | When the trial ends without a paid plan, scoring pauses and events are still accepted through the grace period. If you do not upgrade before grace ends, the account is deactivated, its credentials are disabled, and ingestion stops. |
| Event ingestion rate | 100 requests per second for each API credential. POST /api/events and POST /api/track share the bucket. |
The next request returns HTTP 429 with status rate_limited. |
| Auth API rate | 10 requests per minute per IP, on /api/auth/. |
HTTP 429, with the message "Too many requests. Please try again later." |
| Sensitive API rate | 10 requests per minute per IP, for account actions, team invitation emails, and billing checkout or portal actions. | HTTP 429, with the message "Too many requests. Please try again later." |
| General API rate | 300 requests per minute per IP, for account, identity, device, escalation, team, and dashboard reads, GET /api/events, and the OpenAPI document. |
HTTP 429, with the message "Too many requests. Please try again later." Event ingestion uses the credential bucket above, not this one. |
| Public tools rate |
Signed out: 10 requests per minute and 100 per day, per IP, on /api/tools/.
Signed in: 100 requests per minute per IP.
|
HTTP 429 when the rate is exceeded. Signed-in Authenticity Check calls also spend the monthly operations quota. |
| MCP rate | 60 requests per minute for each signed-in user. A call with no user is limited by IP, at the same rate. OAuth under /mcp/oauth/ is 10 requests per minute per IP. |
HTTP 429, with the message "Too many requests. Please try again later." |
| Request body size | 64 KB for POST /api/events and POST /api/track. 1 MB for other requests. |
A declared Content-Length over the cap is refused with HTTP 413 and the message "Request body exceeds the maximum allowed size." |
| Event id | At most 64 characters, and it can't start with dregs-. |
|
| Score history length | Up to 500 entries. The default is 100, on the API and on MCP. | A higher limit is cut to 500. The call still succeeds. |
| MCP list length | Most list tools default to 25 and return at most 100. Dataset entry lists default to 50 and return at most 500. | A higher limit is cut to the maximum. The call still succeeds. |
| Webhook delivery | 3 attempts. The first retry waits 30 seconds, and the second waits 60 seconds. Connect, write, and read each time out at 10 seconds. | A status other than 2xx, a timeout, or a network failure is recorded and retried. After the third attempt the delivery is marked failed. You can read the attempts on the channel's delivery log. |
| Webhook rate | 600 deliveries per minute and 10,000 per hour, per team. | A delivery over the rate is recorded as failed with "Webhook rate limit exceeded" and follows the same retries. |
| Group types and ids | Up to 10 group types per team. A type normalizes to at most 32 characters. A group id sent on an event is cut at 64 characters. |
On an event, an invalid type or an eleventh type is dropped and the rest of the event
is still saved. POST /api/groups past 10 types is refused with HTTP 400
and the message "A team can use at most 10 group types."
|
| Email to one address | 5 emails per minute to the same address. | The send is refused with "Too many emails sent to this address. Please try again in a minute." This covers confirmation, password reset, invitations, and notification email. |
Related pages
- Pricing: plan prices and the included allowances.
- Events: what an event is, and how to send one.
- REST API: status codes for ingestion, reviews, and paging.
- Webhooks: signatures, headers, and the retry schedule.
- Groups: types, ids, and membership.
- Team management: roles, invitations, and API credentials.
- AI agents: the MCP server.