Feature

Actions on sites

If a browser can do it, Parse can do it — as an API. Log in, fill a form, complete a booking, all from one call.

POST /scraper/{id}/book
curl -X POST .../scraper/{id}/book \ -H "X-API-Key: pmx_..." \ -d '{ "date": "2026-07-01", "party_size": 2 }'
200 · response
{ "confirmed": true, "reservation_id": "r_8f3a", "time": "7:30 PM" }
One request runs the whole flow.
POST = actionsNo browserCaptcha handledSelf-healing
Why

One call instead of a hundred clicks

Most automation tools drive a real browser — booting a headless Chromium, clicking through the DOM, and hoping the page didn't shift. Parse skips the browser. You get a single endpoint that performs the whole flow and returns structured JSON, so a task that used to be a brittle script of clicks becomes one durable HTTP request.

Endpoints

POST endpoints that do things

GET endpoints turn a page into typed JSON. POST endpoints take actions — they submit, log in, book, and trigger — and return whatever the site gives back, including data that only exists after the action runs.

curl -X POST https://api.parse.bot/scraper/{scraper_id}/{action} \
  -H "X-API-Key: pmx_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{ "date": "2026-07-01", "party_size": 2 }'

The exact parameters and a ready-made snippet live on the endpoint's page in the dashboard.

Infrastructure

Hosted, so you never touch infra

Everything that makes actions flaky elsewhere is handled on our side:

  • Proxy rotation and anti-bot handling, hosted — you never configure or pay for infrastructure.
  • Captcha-solving and retries built into every call.
  • The endpoint is continuously health-checked and self-heals when a site changes.
Behind a login?
See Authenticated APIs — Parse logs in, holds the session, and runs your protected actions against it.