Skip to content
Knowledge Base

Train your agent on
your data.

Train intoCHAT on your website, files, Notion pages, text and Q&A pairs. The agent answers from that content and says when it doesn't know.

Website Crawler

Enter your URL and intoCHAT finds up to 500 pages from your sitemap and links. Choose the pages to read, exclude paths, and retrain when your site changes, by hand or automatically every day or week.

Sitemap detection
JavaScript rendering
Exclude paths

Documents (PDF, Word and more)

Upload product manuals, policies or sales playbooks. PDF, Word, PowerPoint, OpenDocument, RTF, EPUB, plain text and Markdown files up to 50 MB each.

Headings and tables kept
Split along paragraphs
Preview and download

Custom API actions

Let the agent call your own HTTP API during a chat to look up live data, such as an order's status or a product's stock.

Bearer, basic or API key auth
Import from a cURL command
Choose which fields the model sees

Spreadsheets (CSV, Excel)

Good for product lists, price lists or branch locations. Their contents become part of the knowledge the agent searches.

CSV, XLS and XLSX
OpenDocument spreadsheets
Searched by keyword and meaning

Source management

Every source shows its status. Edit, preview, download, delete or retrain it, and star the ones that should rank higher.

Star a source for priority
Retrain when content changes
Filter by status or type

Raw Text & Q&A

Add text snippets, or Q&A pairs for answers that must be worded exactly. Q&A answers take priority over your other sources.

Text snippets
Q&A answers come first
Edit at any time

Ready to see it in action?

Create a free account, add your website or a PDF, and chat with your agent in the playground.

Start Building Now