FAQ
About the product
What is CrawlBrain? Is it a chatbot?
CrawlBrain is your website search, upgraded with AI. Visitors type a question and get a clear answer based on your content, with links to the original source pages.
It is not a chatbot. It works where visitors already expect to search: relevant pages appear as they type, and when they submit the question, a direct answer appears above the results. There is no persona, no scripted flow, and no decision tree to build.
Who is it for?
Any website with content people search: online stores, SaaS companies, help centers and support teams, blogs, publishers and WordPress sites, businesses with international visitors, and technical or specialist knowledge bases.
Where do the answers come from?
The answers come from your website and nothing else. The AI reads the relevant slices of your own content and writes the answer from them, so it never draws on the open web or on general knowledge to fill a hole.
If your content does not say it, the answer does not claim it.
Can it make things up in front of my customers?
No. It answers only from your pages and shows the sources on every answer, so a visitor can check them in one click.
When it has nothing to cite, it writes no answer: it tells the visitor the content is not there and logs the question as a content gap.
The realistic risk is a confident answer taken from one of your own outdated pages. The fix is to exclude that page from AI.
Do I need a developer?
No. Paste your URL, then copy one line of code, or install a plugin with no code at all. The widget finds your search bar itself.
How long until it is live?
It is live within minutes. Entering your URL takes a few seconds and adding the widget about 30 seconds, plus 2 to 10 minutes for the first crawl on most sites.
Does it slow down my site?
No. The script loads asynchronously and never blocks rendering.
Content
What does it index?
It indexes every page it can reach from your sitemap and your internal links: products, docs, FAQs, blog posts and policies.
It skips anything behind a login, anything you exclude, anything marked noindex, and non-HTML files.
How often does it re-sync?
It re-syncs automatically every night. Press Sync now when you have published something you want answered today.
Can I add content that is not on my website?
Yes. Snippets are content you write directly, for answers that have no public page. Or import a document in bulk.
Can it read PDFs?
The crawler does not read PDFs, but you can import them as snippets, which extracts the text and splits it at headings. Scanned PDFs need OCR first.
Does it index content behind a login?
No. It indexes what a visitor can reach.
Languages
What languages does it handle?
It answers in the language the visitor asks in, whichever language your pages are written in. The 20 languages you see in the settings are for the widget's own interface: the button, the placeholder and the fallback message come translated into each of them. Keyword search also reads plural and case endings in those 20, including inflected ones like Spanish, Polish, Hungarian and Turkish.
Do I have to translate my content?
No. Your content stays exactly as it is. A visitor asks in Portuguese on your English site and gets a Portuguese answer sourced from your English pages.
How does it know which language to answer in?
It works it out from the visitor's question and their browser, so there is nothing for you to configure.
The widget
Will it replace my existing search bar?
It can. Replace search bar takes over your existing input and expands in place while keeping your styling. You can also run it alongside yours, or as a floating button. See Placements.
Can I change how it looks?
Yes: the theme, fonts, colors, sizes, radius, spacing, icon and every string can be changed. See Appearance.
Can I remove the CrawlBrain branding?
Yes, with Pro. Without it the footer stays.
Can I use it on multiple websites?
You can connect one website on the free tier and three with Pro. Each site keeps its own index and conversation log and gets its own widget, while all of them share one credit balance.
Can it behave differently on different pages?
Yes, with data-* attributes on the script tag, so one agent can use a different placement on each page, or different colors and sizes. See Per-Page Overrides.
Can I open it from my own button?
Yes. Set data-trigger-selector to your button's id, or call window.openWidget() from your own code. See JavaScript API.
Pricing
How does pricing work?
You buy prepaid credit in dollars. Answers, searches, syncing and imports all draw on that one balance, each priced at what it actually costs to run. See Credits and Limits.
What is the difference between searches and answers?
A search shows relevant pages as someone types. It is fast and costs a small fraction of a cent, because no AI writes anything.
An answer is the AI writing a direct response with source links when they submit a question. It uses more credit, depending on the AI model you choose.
That is why the widget searches as people type and only calls the AI on Enter. See Answers and Searches.
Can I try it without paying?
Yes. Create a Free account and get a one-time $5 signup credit, no card required. Connect one website with up to 1,000 pages and test search and AI answers on your own content.
Does credit expire?
No. It sits on the account until it is used.
What is Pro for?
Pro costs $5 every 30 days, and the $5 lands in your credit balance. It adds bulk downloads, the AI model picker, branding removal, the API, and higher page and website ceilings.
Is the $5 for Pro charged on top of my credits?
No. The $5 becomes credit. What you pay for beyond that is the features and the higher ceilings.
What happens if I run out of credit?
The widget switches to free keyword search, so visitors can still find pages. AI answers and semantic search pause until you add credit, and so does the daily sync. Your index and settings stay saved.
Can I cancel any time?
Yes, there is no contract. You keep the Pro features to the end of the period you paid for, and your credit balance and your data are untouched.
Does it work for e-commerce?
Yes. Product questions, shipping, returns, sizing, and compatibility are exactly what people search for before they buy. Most store questions are policy questions, and your policy pages are usually well written.
Data and privacy
What do you store about my visitors?
We store the question, the answer, the sources used and a per-session identifier. Names and email addresses are not stored, and neither are IP addresses.
Visitor labels like SharpFalcon063 are generated per session and are not linked to a person.
How long is history kept?
Conversations and searches are kept for 60 days, and so are sync logs. Your indexed content stays until you delete it.
Export anything you need to keep longer.
Is it GDPR compliant?
Yes. Payments run through Paddle, who handle tax and compliance worldwide.
Can I delete everything?
Yes. Delete a website to remove its index, or delete your account from Account Settings.
Integrations
Is there an API?
Yes, with Pro. See API Overview.
Can I export my content?
Yes, as llms.txt, llms-full.txt, a Markdown bundle, CSV or JSON. llms.txt is free on every account; the rest need Pro. See Exports and llms.txt.
What is llms.txt for?
It gives AI systems a clean map of your site instead of making them guess from your HTML, the way robots.txt does for crawlers. Publish it at your site root so ChatGPT, Claude, Gemini, and Perplexity can quote you accurately.
The company
Who builds CrawlBrain?
CrawlBrain is built by the team behind Squirrly, WP Ghost and other established WordPress and SaaS products. We have spent 16 years building and supporting software for online businesses, and CrawlBrain is built for the same long-term development and support.