Home
Softono
Komorebi

Komorebi

Open source BSD-3-Clause TypeScript
Website GitHub
47
Stars
0
Forks
0
Issues
0
Watchers
1 month
Last Commit
AI Tools Read-it-Later & RSS

About Komorebi

Desktop RSS application built specifically for AI practitioners

Platforms

Web Self-hosted

Languages

TypeScript

Links

Website Source Code
h
Published by
hiwfwang
Visit View Profile

README.md

View on GitHub

Komorebi icon

Komorebi

Komorebi is not a general-purpose RSS reader.
It is a desktop RSS application built specifically for AI practitioners who want to stay current with:

  • frontier industry thinking and commentary
  • newly released AI products and real-world applications
  • state-of-the-art papers and technical research

Overview

Komorebi is an open-source, always-free RSS desktop app for Windows and macOS. It combines a curated AI-focused information network with a reading experience optimized for high-signal sources such as X/Twitter, YouTube, newsletters, and research feeds.

The project is designed to be useful out of the box: users can start with recommended sources immediately, then expand into their own custom feeds and sync services as needed.

Why Komorebi

  • Built for AI professionals: the product focuses on information density, source quality, and workflows relevant to people working in AI.
  • Curated from day one: Komorebi ships with a large set of recommended AI sources, plus guides for adding more.
  • Optimized reading experience: special attention has been given to content presentation, especially for YouTube and social-media-style sources.
  • Optional AI features: AI-powered workflows are available, but entirely opt-in. Users bring their own API keys and model endpoints.
  • Free and open source: the project is intended to remain free to use and open to community contributions.

Core Features

  • AI-focused recommended feeds and onboarding guides
  • Multiple feed presentation modes, including card, list, magazine, and compact views
  • Full desktop reading workflow with unread tracking, starring, search, and article management
  • AI Search for discovering and summarizing relevant articles within a selected time range and topic
  • AI-based article translation
  • Token usage tracking for AI requests
  • YouTube transcript retrieval and transcript-based summarization workflows

Product Site

The product website is available at komorebi-homepage.vercel.app.

Installation

Komorebi is a desktop application for Windows and macOS. Prebuilt packages are available on the GitHub Releases page.

[!IMPORTANT] Network access is a hard requirement for both the website and core in-app content retrieval. If you are in mainland China or another restricted network environment, you should expect to use a VPN or a correctly configured proxy/PAC setup before Komorebi can work reliably.

macOS Note

If macOS shows the message that the app is damaged and should be moved to the Trash, run the following command in Terminal:

sudo xattr -rd com.apple.quarantine /Applications/Komorebi.app

AI Configuration

Komorebi's AI features are optional. To use them, configure your own OpenAI-compatible endpoints, API keys, and model names inside the app settings.

Current AI-related capabilities in the codebase include:

  • chat-model-powered AI Search
  • embedding-based article retrieval and filtering
  • article translation
  • token usage history

Tech Stack

  • Electron
  • React
  • TypeScript
  • Redux
  • Webpack
  • Lovefield

Contributing

Bug reports, suggestions, and pull requests are welcome.

If you run into an issue or have an idea for improving the product, please open an issue on GitHub Issues. Contributions are especially welcome in product polish, source integrations, AI workflows, and reading experience improvements.

Acknowledgements

Special thanks to these open-source projects and resources:

License

BSD-3-Clause