Parallel needed a page behind a login.Browser Use opened it.

Parallel's Task API hands login-protected sources to the Browser Use MCP, which completes the login and fetches the page. The answer comes back with that source cited.

Parallel's published example

Before and after connecting the Browser Use MCP: the report that stopped at NXP's account login, then the completed task with the K66-180 migration details and their source.

Source

Parallel keeps the task. Browser Use reaches the page.

The login stays in the browser session. Parallel gets the source back and finishes the answer.

The Browser Use homepage handoff: Parallel's task reaches toward the logged-in web
Parallel taskLogged-in web

Add Browser Use as an MCP server.

One mcp_servers entry on a Task API run. This is Parallel's sample request for the NXP task.

From Parallel's announcement.

POST /v1/tasks/runsParallel Task API

curl -X POST "https://api.parallel.ai/v1/tasks/runs" \
  -H "x-api-key: $PARALLEL_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
    "input": "Go to https://www.nxp.com/products/K66_180 and extract only the migration-related information for the K66-180 chip, specifically documentation on Migration from Kinetis K Series to MCXNx4x Series.",
    "processor": "ultra",
    "mcp_servers": [{
      "type": "url",
      "url": "https://api.browser-use.com/mcp",
      "name": "browseruse",
      "headers": { "Authorization": "Bearer YOUR_BROWSERUSE_API_KEY" }
    }]
  }'

Connect Browser Use to your Parallel tasks.