Today we're launching /interact — a new endpoint that turns any Firecrawl scrape session into a live browser you can control.

• /interact — scrape a page and immediately take actions in it: click buttons, fill forms, navigate, and extract data that's only reachable through interaction
• Natural language actions — describe what to do in plain English and Firecrawl handles the clicks, typing, and navigation
• Code execution — write Node.js, Python, or Bash to control the browser with full Playwright access
• Live view — embed a read-only or interactive stream of the browser in your own UI
• Persistent sessions — state (cookies, localStorage, login) carries over between interactions and across scrapes

Most web data you care about is behind an action. A search form. A login. A "load more" button. A filter dropdown.

Until now, Firecrawl gave you clean data from a URL — but what the page showed after an interaction was out of reach.

With /interact, you scrape a page and stay in that session. Then you tell it what to do next.

The workflow is three steps:

1. Scrape — POST /v2/scrape returns a scrapeId along with the page's content
2. Interact — POST /v2/scrape/{scrapeId}/interact takes actions using a prompt or code
3. Stop — DELETE /v2/scrape/{scrapeId}/interact ends the session

Sessions stay alive for up to 10 minutes by default (with a 5-minute inactivity timeout), and you can chain as many interaction calls as you need within that window.

Describe what you want in plain English. No code required.

# CLI
firecrawl interact "Search for iPhone 16 Pro Max"
firecrawl interact "Click on the first result and tell me the price"

Each prompt should be a single, clear task. The agent handles the mechanics — finding elements, clicking, typing, waiting for results.

For precise control, pass Playwright code directly using the Firecrawl CLI. Node.js, Python, and Bash are all supported. The page object is pre-connected to the live browser session — no setup needed.

# Uses the last scrape automatically
firecrawl interact -c "
  await page.click('#next-page');
  await page.waitForLoadState('networkidle');
  const title = await page.title();
  const content = await page.\$eval('.article-body', el => el.textContent);
  JSON.stringify({ title, content });
"
 
# Or pass a scrape ID explicitly
# firecrawl interact <scrape-id> -c "await page.title()"

Every /interact response includes two URLs:

• liveViewUrl — a read-only stream, useful for monitoring or embedding in dashboards
• interactiveLiveViewUrl — a fully interactive stream where a human can take over the browser directly

This makes it easy to build tools that combine automated scraping with human-in-the-loop review.

Need to stay logged in across scrapes? Named profiles save browser state — cookies, localStorage, sessions — so you don't re-authenticate on every call.

# CLI: scrape the login page, log in, then stop (state is saved)
firecrawl scrape https://app.example.com/login --profile my-app
firecrawl interact "Fill in user@example.com and password, then click Login"
firecrawl interact stop
 
# Session 2: scrape with the same profile — already logged in
firecrawl scrape https://app.example.com/dashboard --profile my-app
firecrawl interact "Extract the dashboard data"
firecrawl interact stop

The profile persists until you delete it, making recurring scrapes on authenticated pages much simpler.

Navigate paginated results — click "next page" in a loop and extract data at each step, all in one session.

Fill and submit forms — trigger searches, apply filters, or submit queries to get dynamic results back as clean data.

Monitor live pages — embed the liveViewUrl in an internal tool so your team can watch a scrape unfold in real time.

Automate multi-step workflows — chain prompts or code blocks to complete a full task: search, filter, click, extract.

• Read the docs: docs.firecrawl.dev/features/interact
• Join the community: Share what you build in our Discord.