How to train an AI agent on your website

Ground a Solveo AI agent on your own site by crawling a URL, uploading documents, or adding Q&A entries in the knowledge base.

Steps

Do this in order

  1. Open the knowledge base

    In the dashboard, open your agent and go to the knowledge base. This is where uploaded files, crawled pages, and Q&A entries are indexed.

  2. Crawl your website

    Choose Website Crawl, paste a root URL, and set the maximum number of pages. The crawler follows that site and shows progress while pages are indexed.

  3. Add documents or Q&A for precise answers

    Upload PDF, DOCX, or TXT files, or add Q&A entries for FAQs and policy statements. Q&A entries take priority over document-derived answers.

  4. Re-crawl when the site changes

    Run the website crawl again whenever docs or product pages change so the agent keeps answering from current content.

Solveo AI retrieves answers from your knowledge base rather than guessing. A website crawl is the fastest way to ground an agent on public docs or a help center. Uploads and Q&A entries cover policies and files that are not on the public site.

Full source types and limits are in Knowledge Base.

FAQ

Common questions

What file types can I upload to the knowledge base?
PDF, DOCX, DOC, TXT, and MD, up to 50 MB per file. Files are chunked automatically for retrieval.
Can the agent learn from a JavaScript-rendered site?
Yes. Website crawl includes a stealth mode for modern JavaScript-rendered sites. Re-crawl after you publish updates.

Related docs

Get started

Put the steps into practice

Start free, no credit card required.