Helicone · Rate Limits

Helicone Rate Limits

Helicone publishes per-tier ingestion and API call ceilings on its pricing page rather than a dedicated rate-limits doc. The AI Gateway also enforces strict caps on traffic to non-approved target domains. Limits are enforced per organization / API key. Headers and response codes for throttling are not publicly documented; consumers should rely on standard 429 handling and exponential backoff.

11 Limits Throttle: 429 Quota: 429
AI GatewaysAI MonitoringGatewaysLLM ObservabilityLLM RoutingPrompt ManagementRate LimitingQuotasThrottling

Limits

Hobby Tier Log Ingestion organization
logs_per_minute · minute
10
Sustained log ingestion ceiling for the free Hobby tier.
Hobby Tier API Calls organization
requests_per_minute · minute
10
API call ceiling against Helicone REST APIs on Hobby.
Hobby Tier Monthly Requests organization
requests_per_month · month
10000
Logged-request quota; further requests roll into usage-based pricing on paid tiers.
Pro Tier Log Ingestion organization
logs_per_minute · minute
1000
Pro Tier API Calls organization
requests_per_minute · minute
60
Team Tier Log Ingestion organization
logs_per_minute · minute
15000
Team Tier API Calls organization
requests_per_minute · minute
1000
Enterprise Tier Log Ingestion contract
logs_per_minute · minute
30000
Negotiable beyond the published 30k logs/min baseline.
Gateway - Unapproved Domain Daily Quota api-key
requests_per_day · day
10000
When routing through Helicone-Target-Url to a non-approved domain, requests are capped at 10,000 per day.
Gateway - Unapproved Domain Rate api-key
requests_per_second · second
1
When routing through Helicone-Target-Url to a non-approved domain, requests are capped at 1 per second.
Gateway - Approved Provider Domain api-key
varies
inherits underlying provider limits (OpenAI, Anthropic, Google, etc.)
For approved provider targets the gateway does not enforce its own limits; the upstream provider's rate limits and 429 behavior apply.

Policies

Backoff Strategy
Helicone does not document per-tier headers; clients should implement exponential backoff with jitter on any 429 or 503 response and honor a Retry-After header if present.
Burst Handling
Sustained ingestion above the per-tier logs/min ceiling will be throttled. Short bursts may be tolerated, but no published burst ratio is guaranteed.
Quota Reset
Monthly request quotas (10,000 base across tiers) reset on the billing-cycle boundary; additional usage is metered through the pricing calculator.
Approved Domain Policy
Traffic through the AI Gateway to approved provider hostnames inherits the upstream provider's rate limits. Traffic to unapproved hostnames is capped at 10,000/day and 1/second per API key.
Tier Upgrade
To raise ingestion or API call ceilings, upgrade tier (Hobby - Pro - Team - Enterprise) or contact sales for an Enterprise contract.

Sources