Running Extractions - Extralt Docs

A run is an extraction job. You select a robot, provide start URLs, and Extralt crawls the site to produce captures.

Creating a run

curl -s -X POST "https://api.extralt.com/runs" \
  -H "Authorization: Bearer $EXTRALT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "robotId": "your-robot-id",
    "urls": [
      "https://example-store.com/products/sneakers",
      "https://example-store.com/products/boots"
    ],
    "budget": 100
  }' | jq

Parameters

Parameter	Required	Description
`robotId`	Yes	The robot to use for extraction
`urls`	Yes	Start URLs to crawl
`budget`	No	Maximum number of URLs to extract. Each URL costs 1 credit.

Run lifecycle

Status	Description
`pending`	Run is queued for execution
`running`	Actively crawling and extracting
`completed`	Finished successfully
`failed`	Encountered an unrecoverable error
`stopped`	Manually stopped

Monitoring a run

curl -s "https://api.extralt.com/runs/$RUN_ID" \
  -H "Authorization: Bearer $EXTRALT_API_KEY" | jq

You can also monitor runs in the dashboard.

Concurrent run limits

Plan	Concurrent runs
Start	1
Scale	Up to 10

If you exceed your concurrent run limit, the run will be queued until a slot opens.

Downloading data

After a run completes, you can download all capture data as a .jsonl.lz4 file via the API. The download endpoint returns a signed URL valid for 10 minutes.

curl -s "https://api.extralt.com/runs/$RUN_ID/download" \
  -H "Authorization: Bearer $EXTRALT_API_KEY" | jq '.url'

Dashboard download is coming soon. For now, use the API to download capture data.

What's next

Working with Captures -- access individual captures
Common Patterns -- polling runs until completion
Credits & Billing -- understand credit usage