Better Data

Hosted

Rate Limiting

Edit this page

Better Data Cloud provides sophisticated rate limiting to protect your commerce infrastructure from being overwhelmed by AI agents or malicious actors.

Multi-Layer Protection

We apply rate limits at three different levels:

1. Account Level

Enforced based on your Subscription Tier. This protects the Better Data infrastructure.

2. Provider Level (Guardrails)

You can set limits on specific LLM providers. For example, you might want to allow Claude to make 500 calls/hr but limit OpenAI to 200 calls/hr.

3. User / Session Level

Prevent a single user or bot from monopolizing your resources.

  • Default: 20 calls per session per minute.
  • Customizable: You can increase or decrease this in the dashboard.

Response Headers

Every request to the Hosted Gateway includes headers indicating your current status:

1X-RateLimit-Limit: 1000
2X-RateLimit-Remaining: 950
3X-RateLimit-Reset: 1704456000

Handling 429 Errors

When a limit is reached, the gateway returns:

1{
2  "error": "rate_limit_exceeded",
3  "message": "You have exceeded your hourly quota. Upgrade to a higher tier for more capacity.",
4  "retry_after": 300
5}

We recommend that your client application or LLM adapter handles these errors gracefully by notifying the user or retrying after the retry_after duration.