Rate Limits

Rate limits protect platform stability and prevent one account from consuming too much capacity in a short time. Limits may vary by account, token, model, or benefit tier.

Common Terms

• RPM: requests per minute.
• RPD: requests per day.
• Concurrency: the number of requests being processed at the same time.
• Global limit: the total limit shared by multiple tokens in an account.
• Token limit: an additional limit assigned to one API key.

Client Design

Production applications should do three things:

1. Control concurrency instead of sending every request at once.
2. Use exponential backoff for 429 responses, temporary upstream errors, and network errors.
3. Record the model, status code, and request ID for failed requests.

Exponential Backoff Example

Retry 1: wait 1 second
Retry 2: wait 2 seconds
Retry 3: wait 4 seconds
Retry 4: wait 8 seconds

If repeated retries still fail, stop retrying and escalate the error for manual handling.

Increase Capacity

If your use case requires a higher rate or more concurrency:

• Check Benefit Tiers for your current level and capacity rules.
• Set lower token limits for test scripts so they do not affect production services.
• Contact the DDTokens team to discuss dedicated quotas or enterprise plans.

Exact RPM, RPD, and concurrency limits are determined by the dashboard and your account rules.