Timeouts · Docs | ScreenshotDude

Server limit: 60 seconds

Synchronous requests have a 60-second server-side limit, which covers loading the page, waiting (wait_until, wait_for_selector, delay), scrolling and capturing. If the render doesn't finish in time the API returns 500 and you're not billed.

Most requests finish in one to a few seconds. Long ones are usually caused by:

wait_until=networkidle0 on pages with constant background requests (analytics, websockets);
large delay values (max 30 s);
very long full-page captures with full_page_scroll;
slow proxies.

Set your HTTP client timeout slightly above the server limit (for example 65 seconds) so you receive the API's error rather than a client-side abort.

Use is_async for long work

With is_async=true the API immediately returns 202 with a job_id. The job then runs without the 60-second HTTP constraint of your client. Either poll /v1/jobs/{id} (free) or pass a webhook_url to be notified. See Async jobs.

curl -G "http://localhost:4000/v1/capture" \
  --data-urlencode "url=https://heavy.example" \
  -d full_page=true -d full_page_scroll=true -d is_async=true \
  -H "x-api-key: $SNAPLENS_API_KEY"
# → {"success":true,"job_id":"job_…","status":"Queued"}

Platform tips

Serverless platforms often stop your function before a slow capture finishes.

Platform	Default / max duration	Recommendation
Vercel Functions	Short default; configurable `maxDuration` depending on plan	Raise `maxDuration` to at least 65 s for sync calls, or use `is_async` + webhook.
AWS Lambda	3 s default, up to 15 min	Set the function timeout to ≥ 65 s. Behind API Gateway (29 s limit), use `is_async`.
Netlify Functions	10 s default, up to 26 s synchronous	Use `is_async` + webhook, or Background Functions.
Cloudflare Workers	Generous wall-clock time while waiting on fetch, but subrequest limits apply	Sync calls usually work; for batches use Queues + `is_async`.
Make.com	HTTP module timeout is configurable (default 40 s)	Raise the timeout to 60–65 s or use `is_async` with a webhook trigger scenario.
Zapier	Actions time out after about 30 s	Use `is_async=true` with a "Catch Hook" Zap as `webhook_url`.

Limits change over time; check your platform's documentation for current values.

Retries and idempotency

A timed-out synchronous request is never billed, so retrying is safe. For async jobs, store the job_id and poll it instead of re-submitting.