Rate Limiting

The Pixee PIM API enforces per-endpoint rate limits to ensure fair usage and platform stability. Limits are tracked in Redis and persist across API instances.

Rate limit tiers

Endpoints are grouped into tiers by category, each with its own limit and sliding window.

Auth

Tier	Limit	Endpoints
`AUTH_LOGIN`	5 / 15 min	`POST /auth/login`
`AUTH_PASSWORD_RESET`	3 / hour	`POST /auth/forgot-password`
`AUTH_REFRESH`	30 / min	`POST /auth/refresh`

AUTH_LOGIN uses a 15-minute window — not 1 minute. After 5 failed login attempts the endpoint is blocked for the remainder of that window.

Read

Tier	Limit	Endpoints
`READ_PUBLIC`	2 000 / min	Health checks, public routes
`READ_STANDARD`	1 000 / min	GET endpoints (products, imports, etc.)
`READ_ADMIN`	200 / min	Admin endpoints
`READ_SENSITIVE`	50 / min	Analytics, KPIs, reports

Write

Tier	Limit	Endpoints
`WRITE_STANDARD`	100 / min	POST / PUT / PATCH / DELETE
`WRITE_CONFIG`	30 / min	Configuration endpoints
`WRITE_UPLOAD`	30 / min	File upload endpoints
`WRITE_USER_MGMT`	20 / min	User management

Bulk

Tier	Limit	Endpoints
`BULK_UPDATE`	30 / min	Bulk attribute updates
`BULK_EXPORT`	20 / min	Export batch jobs
`BULK_CREATE`	20 / min	Bulk product creation
`BULK_IMPORT`	10 / min	Import batch jobs
`BULK_DELETE`	10 / min	Bulk deletes

AI

Tier	Limit	Endpoints
`AI_SINGLE`	10 / min	Individual AI enrichment
`AI_BATCH`	5 / min	Batch AI enrichment

Export & utilities

Tier	Limit	Endpoints
`PROGRESS_POLL`	600 / min	Job progress polling
`HEALTH_CHECK`	600 / min	Health endpoints
`EXPORT_DOWNLOAD`	100 / min	File downloads
`EXPORT_STANDARD`	50 / min	Standard export jobs
`EXPORT_STREAM`	10 / min	Streaming exports

Rate limit headers

X-RateLimit-* headers are injected only on authenticated endpoints accessed via API key. They are not present on every response.

Header	Description
`X-RateLimit-Limit`	Maximum requests allowed in the current window
`X-RateLimit-Remaining`	Requests remaining in the current window
`X-RateLimit-Reset`	Unix timestamp when the window resets

Example response headers (API key endpoints)

X-RateLimit-Limit: 1000
X-RateLimit-Remaining: 987
X-RateLimit-Reset: 1747220460

Handling rate limit errors

When you exceed the limit, the API returns 429 Too Many Requests:

429 Response

{
  "error": "rate_limit_exceeded",
  "message": "Rate limit exceeded",
  "detail": "5 per 15 minutes",
  "retry_after": "Please wait before making more requests"
}

The retry_after field is a human-readable string, not an integer. To determine the exact wait time, parse the detail field (e.g. "5 per 15 minutes") or implement exponential backoff (see below).

Retry strategy

We recommend exponential backoff with jitter:

Exponential backoff (Python)

import time
import random
import httpx

def call_with_backoff(url, headers, max_retries=5):
    for attempt in range(max_retries):
        response = httpx.get(url, headers=headers)
        if response.status_code == 429:
            wait = (2 ** attempt) + random.uniform(0, 1)
            time.sleep(wait)
            continue
        return response
    raise Exception("Max retries exceeded")

For bulk operations (imports, exports, batch AI enrichment), schedule jobs during off-peak hours. BULK_IMPORT, BULK_DELETE, and AI_BATCH have the strictest limits (5–10 requests per minute).