Data

Overview

You can connect your bot to external data. It can be a web page, a document or an entire website.

WARNING

You need to set an OpenAI API key in account settings in order to analyze data. Kenja uses the text-embedding-3-small model to create sematic index for the data.

Website

Kenja will automatically crawl the website and extract the contents as markdown for you. By default it fetches every page on the website starting from the URL you entered, but you can specify what pages you want to include:

/
/about
/blog/**

Notice that you can use * to match any characters, and ** to match any characters including /. This syntax is powered by micromatch.

File

You can upload .txt, .md, .pdf files.

Notion

Planned.