How it works
Firecrawl turns web pages into content your app can read and organize. Connect it to collect information from a page, discover pages on a site, or search the web. App visitors don’t need their own Firecrawl accounts; your app uses the key you connect. By connecting Firecrawl, you can:- Read the main text from a web page
- List pages discovered on a website
- Search the web for relevant pages
- Collect content from multiple pages in a crawl
- Extract specific details, such as a product name and price
- Request summaries of pages found in a search
How to integrate Firecrawl into your app
1
Connect to Firecrawl
Ask Zite to connect your app to Firecrawl. Alternatively, select + in the chat box, then Integrations → Firecrawl.Create a key in your Firecrawl account and enter it in API Key. If you’ve connected before, choose a saved account or Connect new account.
2
Describe your use case
Tell Zite which pages to read, what information to collect, and where to show or save it. For example, “Let our team enter a supplier’s product page and extract the name, description, and price for review.”For a whole site, ask it to list the discovered pages first so you can choose what to collect.
3
Test and publish
Test one page first, then a small group of pages, and compare the results with the originals. For longer crawls, ask Zite to show progress and any pages that failed.Select Publish in the upper-right corner when the feature works as expected.
Common use cases
With a Firecrawl connection, you can create apps like:- Supplier Research: Collect product details for your team to review
- Page Inventories: List discovered pages before a website content review
- Research Libraries: Save relevant page text with its source link
- Market Briefings: Search for relevant news and summarize selected results
- Product Comparisons: Extract names and prices from selected product pages
Important notes
- Your own key: Firecrawl requires your own key; there is no Use Zite’s Account option.
- Usage and limits: Activity from all app visitors uses your Firecrawl allowance. Set a page limit for each crawl and check your balance before running a large job.
- Choose the smallest task: Read one page when you need its content. Listing a site’s pages doesn’t collect the text of every page.
- Longer crawls: A crawl can take minutes. Ask for a progress view that checks the job rather than starting it again whenever the page refreshes.
- Incomplete results: Check extracted details against the source and review failed pages. Don’t assume a crawl found every page or that a search result already contains the full article.
- Failed requests: Check that the key is valid and the account has credits. An exhausted balance needs attention in Firecrawl; retrying alone won’t fix it.
Need help with this integration? Contact support@zite.com.