Credits and limits
Understand pay-as-you-go pricing, free tests, concurrency, and usage tiers.
Berrycrawl uses flat pay-as-you-go credits. There is no subscription commitment.
- $1 buys 1,000 credits.
- The minimum top-up is $10.
- New workspaces receive 100 credits once.
- Paid credits do not expire.
berrycrawl.comis a free test target.
A successful uncached scrape costs 1 credit, including any proxy escalation needed to retrieve the page. Screenshots, PDFs, JSON extraction, and summaries have explicit format add-ons. Failed, blocked, and cached ordinary page responses are free. The response reports the amount consumed.
Every successful YouTube transcript response costs 5 credits, whether you use POST /youtube/transcript or the compatible POST /scrape path. Missing captions and failed requests
are not charged.
Brand profiles cost 10 credits. Brand retrieval is billed per successful API call, including cache
hits, because the caller receives the same complete profile either way. Set refresh: true only
when you need the website read again.
Web search costs 2 credits per successful request, equivalent to $2 per 1,000 searches. Search uses Parallel Turbo mode and returns up to 10 results with excerpts. Search never scrapes result pages; use the scrape API separately for selected URLs that need full content.
Usage tiers
Lifetime top-up spend raises rate and concurrency limits automatically. Request rate, active scrapes, and queued scrapes are separate limits: a short burst can wait for an active slot without being rejected.
| Tier | Lifetime spend | Requests/minute | Active scrapes | Queued scrapes |
|---|---|---|---|---|
| Free | $0 | 30 | 3 | 25 |
| Tier 1 | $10 | 120 | 10 | 100 |
| Tier 2 | $100 | 300 | 30 | 250 |
| Tier 3 | $500 | 800 | 80 | 500 |
| Tier 4 | $2,000 | 2,000 | 200 | 1,000 |
Free-email signups may receive a lower Free-tier request rate. Enterprise limits are assigned manually.
Handling limits
POST /scrape waits in a bounded, distributed FIFO queue when all active slots are occupied. Queue
time counts against the request's timeout. The response metadata reports queueWaitMs,
queuePositionAtAdmission, concurrencyLimit, and queueLimit. Call GET /account to inspect
your current credits, plan, limits, and queue usage.
A 429 means the per-minute request rate was exceeded, the scrape queue was full, or the caller's
queue wait timed out. Retry it with exponential backoff and jitter. For many unrelated URLs, send
bounded parallel POST /scrape requests within the limits returned by GET /account. Do not retry
validation errors.