Retell AI · Rate Limits

Retell Rate Limits

Retell AI's primary throttling mechanism is call concurrency - the maximum number of simultaneous calls an organization can run. Each org has a base (free) concurrency limit plus optionally purchased concurrency, and an optional burst mode. When concurrency is exhausted, new outbound calls (including batch-call jobs) queue or are rejected until a slot frees up. Current and maximum concurrency are readable via GET /get-concurrency. The management REST endpoints (agents, LLMs, phone numbers, calls) are also subject to standard per-account request rate limits; exact per-endpoint RPS values are not published and are not reconciled in this artifact.

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

It captures 5 rate-limit definitions, measuring concurrent_calls and requests.

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

Tagged areas include AI, Voice, Voice Agents, Telephony, and Rate Limiting.

5 Limits Throttle: 429
AIVoiceVoice AgentsTelephonyRate LimitingConcurrencyQuotasThrottling

Limits

Call Concurrency (Base) account
concurrent_calls
20
Default free concurrency allotment per organization; readable as base_concurrency.
Call Concurrency (Purchased) account
concurrent_calls
see provider documentation
Additional concurrent-call slots purchased on top of the base allotment; readable as purchased_concurrency.
Concurrency Burst account
concurrent_calls
see provider documentation
Optional burst mode allowing a temporary higher concurrency ceiling (concurrency_burst_limit) when enabled.
Reserved Inbound Concurrency account
concurrent_calls
see provider documentation
Portion of concurrency reserved specifically for inbound calls so campaigns do not starve incoming traffic.
Management API Requests account
requests
see provider documentation
Standard per-account request limits on REST management endpoints (agents, LLMs, phone numbers, calls); not publicly enumerated.

Policies

Concurrency-Gated Calling
Outbound and batch calls consume concurrency slots; when the limit is reached, jobs queue until slots free.
Tiered Limits
Concurrency raises with plan tier and via purchased slots or Enterprise agreements.
Backoff Strategy
Clients should implement exponential backoff with jitter on 429 responses and honor Retry-After where present.

Sources