Pay for what returns.
Every plan includes the same retailer plugins, schema, and SDKs, and draws from a pool of credits. Easy pages cost 1 credit, hard ones more, cache hits are free. Upgrade or downgrade anytime.
- MCP + REST API key
- Community support
- No credit card
- Standard rate limits
- Overage ~$8 per 1k credits
- Batch endpoint
- Email support
- Higher rate limits
- Overage ~$4 per 1k credits
- Price-history feeds
- Priority queue + SLA
- Slack support
- Dedicated throughput
- Custom retailer plugins
- Contractual SLA
- Invoice billing
You pay for success, by how hard the page fought back.
Most pages we reach directly — 1 credit. Pages protected by anti-bot defenses need heavier machinery, so they cost more. We always try the cheapest path first and only escalate when a site forces us to, and you are charged solely for the tier that actually returned your data. The same URL may cost different amounts on different days — pricing follows what each request takes to succeed. Cache hits and failed extractions are always free. Every response reports its exact cost in meta.credits.
| How it was served | Credits |
|---|---|
| Cache hit | 0 |
| Direct fetch | 1 |
| Residential / mobile proxy | 3 |
| JS render | 5 |
| Premium proxy | 8 |
| Unlocker | 20 |
Frequently asked.
What is a credit?+
Credits are the metering unit. Each successful extraction draws credits based on how the page was served: a direct fetch is 1 credit, a residential-proxy page 3, JS render 5, a premium-proxy page 8, and the toughest anti-bot targets up to 20. Your plan includes a monthly credit pool.
Why do some extractions cost more than others?+
Because some pages are harder to reach. A site we can fetch directly is 1 credit; sites guarded by anti-bot defenses need residential proxies, a real browser, or an unlocker to get through, which cost more. We always use the cheapest route that works and charge only for the tier that actually returned your data — so the same URL can cost different amounts on different requests.
What counts as billable?+
Only successful extractions where Cartpie returns a real product object draw credits. Blocked pages, failed extractions, and non-product URLs are free — they never draw credits or appear on your invoice.
What about cache hits?+
Cache hits are free — 0 credits. If you re-request a URL within your freshness window, we return cached data without drawing from your pool.
How do I know what I was charged?+
Every extract response includes meta.credits — the exact credits that call drew. Your remaining balance and usage are in the dashboard and at GET /v1/usage.
What are the rate limits?+
Free: 60 req/min. Indie: 300 req/min. Business: 1,200 req/min. Anchor: custom. All limits are per-account, not per-key.
Do you store the extracted data?+
Only briefly, for caching within your chosen freshness window. We don't sell or share your extraction history.
Which retailers are supported?+
20+ tuned plugins covering most major US, UK, and EU retailers, plus a fallback extractor for the long tail. Full list in /docs.
Do you offer refunds?+
Within 14 days of upgrading, full refund for any reason. After that, prorated on a case-by-case basis.