Knoll is a browser extension for adding external knowledge to AI chatbots (ChatGPT/Claude).
The project is open-source and ad-free.
Knoll has a rich set of features supporting for both adding the information you want and then integrating it into your model's context when relevant.
- ✂️ Clip any text on the Internet to store
- 🔗 Share the text you store with friends
- 📄 Use any Google Doc or Markdown file you have as a knowledge
- 🌐 Import existing knowledge repositories that other users have shared
- 🖥️ Directly integrated into the default chat interface
Knoll is directly integrated with the following AI services.
- ChatGPT
- Claude
You want to make Knoll even better? There are lots of ways you can contribute even if you cannot code yourself:
- Share this project. The simplest way of contributing is by sharing this project with your friends and family.
- Report Issues. If you see some parts of the app not working as expected, if you want to give any kind of feedback or if you just have a question you can submit an issue on the issue page of this project or using our bug report form.
- Write Code. And then there is the standard way of contributing to an open-source project. Please feel free to submit proposals directly in the form of a PR or Issue.
src
- main source.contentScript
- scripts and components to be injected ascontent_script
background
- scripts for background.styles
- styles shared in popup and options pagemanifest.ts
- manifest for the extension.
extension
- extension package root.assets
- static assets.dist
- built files, also serve stub entry for Vite on development.
scripts
- development and bundling helper scripts.
To run the extension, you will need to have our two servers running as well:
pnpm run dev
Then load extension in browser with the extension/
folder.
To build the extension, run
pnpm run build
And then pack files under extension
, you can upload extension.crx
or extension.xpi
to appropriate extension store.
This repo was made based on https://github.com/antfu/vitesse-webext and https://github.com/StanfordHCI/FeedMonitor