Web Scraping Results API - Dexi Robots, Runs and Datasets

Export what your own Dexi web scraping robots have already collected: every scraped row of a run's latest result or of any past execution, the execution history of a run, usage statistics, and dataset rows, as clean rows ready for a warehouse, a spreadsheet or the next step of a pipeline.

What it collects

• Scraped rows: the latest result of a run (from its latest successful execution by default) or the result of any execution by id, one row per scraped row with every field under data, keyed by the robot's own column names.
• Execution history: every execution of a run, newest first, with its state (OK, failed, stopped, running, pending, queued), start and finish time, optionally within a date range.
• Execution statistics: page visits, requests, time and traffic used, current, failed and total results, concurrency and who started it.
• Dataset rows: the rows collected into a dataset, with the provider's own query filters (equals, in, between, greater than and more) and nested fields by dot path.
• Robots and runs: every robot in the account and every run (saved configuration), with ids and names, which is where the ids for everything else come from. Nested robot definitions and any field that looks like a credential are left out.
• Read only, pay per result, bring your own key.

Input

Field	What it does
What to read	Robots (default, needs nothing else), runs, executions of a run, latest result of a run, result of an execution, execution statistics or dataset rows.
Ids	One per line: run ids, execution ids or dataset ids, depending on the service. Robot ids optionally narrow runs.
Latest result from executions in state	Latest result: OK by default, or any state.
From date / To date	Executions: the date range.
Dataset query	Dataset rows: the provider's JSON query.
Maximum results	Row cap for the run.
Requests per minute	Pacing for calls to the provider.
Account id	Your account id, as a secret input.
API key	Your own API key, as a secret input.

FAQ

What is this web scraping results API used for?

Getting data out of scraping robots that already run on a schedule and into the systems that use it. A pricing team pulls the latest competitor price scrape every morning into BigQuery or Snowflake. An analyst exports last month's executions of one run to see which ones failed. An operations team checks page visits and traffic per execution to control scraping costs. A product team reads a dataset of collected listings filtered by category and price.

Which data source does this actor read?

The Dexi API at api.dexi.io, through the read routes documented on its public developer portal: GET /robots, GET /runs, GET /runs/{runId}/executions, GET /runs/{runId}/latest/result, GET /executions/{executionId}/result, GET /executions/{executionId}/stats and POST /datasets/{dataSetId}/rows, which is a query sent as a POST and writes nothing.

Do I need an API key?

Yes. This actor is bring-your-own-key and never ships one. Your account id and API key are both shown on the API page of your Dexi account. Paste both into the input, or paste accountId:apiKey into the API key field, or set them once as the DATA_API_ACCOUNT_ID and DATA_API_KEY environment secrets. The actor computes the documented access key (MD5 of account id and API key) itself, so the key is never sent. Missing or refused credentials end the run cleanly with a message saying which it was.

How do I find run ids and execution ids?

Run the robots service first, then the runs service (optionally with robot ids) to get run ids. The executions service lists a run's executions with their ids and states. The latest result service needs only a run id, which is usually all you need.

Can this actor start a robot or change anything?

No. It never starts, stops or retries an execution, and no create, update or delete route is wired. It reads what your robots have already collected.

What happens when there is nothing for an id?

It becomes its own row with found: false and a note: a run with no successful execution yet, an execution id the account cannot see, or a dataset query that matches nothing. Those rows are never charged.

How is it priced?

Pay per result: one flat price per record returned, whether a scraped row, a dataset row, an execution, a statistics row, a run or a robot. Rows for ids that produced nothing are free. Your Dexi subscription applies separately.

Example output

{
  "service": "latestResult",
  "serviceLabel": "Latest result of a run",
  "requested": "b64a8397-2159-4466-a907-cc615d6530de",
  "found": true,
  "runId": "b64a8397-2159-4466-a907-cc615d6530de",
  "executionId": null,
  "rowIndex": 0,
  "data": {
    "product_name": "Trail Running Shoe 1042",
    "price": "89.99",
    "currency": "EUR",
    "in_stock": "true",
    "url": "https://shop.example.com/products/trail-running-shoe-1042"
  },
  "retrievedAt": "2026-09-28T09:14:52.118Z",
  "note": null
}

Values are illustrative; the columns under data are whatever your robot extracts.

Keyword map

web scraping API, scraping results export, web data extraction API, scraper output to database, scheduled scraping export, competitor price scraping data, scraping execution history, scraping usage statistics, dataset export API, Dexi API, dexi.io robots, web extraction robots.