Agenta · Rate Limits

Agenta Rate Limits

Agenta does not publish explicit numeric per-endpoint request rate limits for its cloud REST API. Practical usage is bounded by plan quotas - included monthly traces (5k Hobby, 10k Pro, 1M Business) with $5 per additional 10k trace overage, monthly evaluation allowances (20 on Hobby, unlimited on Pro and above), seat counts, and data retention windows - rather than by RPM/TPM throttling. Self-hosted deployments are bounded only by the operator's own infrastructure. Specific per-endpoint throttling values are not reconciled in this artifact.

Agenta Rate Limits is the machine-readable rate-limit profile for Agenta on the APIs.io network, conforming to the API Commons Rate Limits specification.

It captures 5 rate-limit definitions, measuring traces, evaluations, seats, days, and requests.

The profile also includes 3 backoff/retry policies defined and response codes documented for throttled.

Tagged areas include AI, LLMOps, Prompt Management, LLM Evaluation, and Observability.

5 Limits Throttle: 429
AILLMOpsPrompt ManagementLLM EvaluationObservabilityRate LimitingQuotasThrottling

Limits

Included Traces account
traces
plan-based (5k Hobby / 10k Pro / 1M Business per month)
Observability traces included per month; overage billed at $5 per additional 10k on paid plans.
Included Evaluations account
evaluations
20/month Hobby; unlimited Pro and above
Monthly evaluation runs allowance; varies by plan.
Seats account
seats
2 Hobby / up to 10 Pro / unlimited Business
Workspace member seats; not a request rate limit.
Data Retention account
days
30 Hobby / 90 Pro / 365 Business
Trace and evaluation data retention window per plan.
Per-Endpoint Request Rate account
requests
see provider documentation
No explicit numeric RPM/TPM published for the REST or OTLP endpoints as of the review date.

Policies

Quota-Based Usage
Cloud usage is bounded by plan quotas (traces, evaluations, seats, retention) rather than published RPM throttling.
Self-Hosting Removes Platform Quotas
MIT-licensed self-hosted deployments are limited only by the operator's own infrastructure and LLM provider limits.
Backoff Strategy
Clients should implement exponential backoff with jitter and honor any Retry-After header on 429 responses.

Sources