Karakeep is a self-hostable bookmark manager that stores links, notes, and images with AI-based automatic tagging and full text search.
The tool addresses the problem of organizing and retrieving saved content at scale. It lets users capture bookmarks, notes, and images through browser extensions, mobile apps, or direct uploads, then automatically fetches metadata like titles and descriptions for links. The system uses large language models to generate tags and summaries, with support for local models via Ollama. Search combines full text indexing with semantic search to help users find saved content by meaning rather than exact keywords. The project also includes OCR for extracting text from images, full page archival to prevent link rot, and video archiving capabilities.
Karakeep suits developers and information workers who want to self-host their bookmark infrastructure rather than rely on third-party services. It works well for teams collaborating on shared bookmark lists and for users who want to integrate bookmarking into automated workflows through its REST API and CLI. The tool imports existing bookmarks from Chrome, Pocket, Linkwarden, Omnivore, and Tab Session Manager, and can sync with browser bookmarks via Floccus. Mobile apps for iOS and Android provide offline reading and quick capture on the go.
The project maintains active continuous integration with passing test workflows. Development activity shows consistent engagement across multiple platforms and integrations, with browser extensions for Chrome, Firefox, and Safari alongside native mobile applications. The codebase spans TypeScript across web and mobile implementations, and the project supports internationalization through community translation contributions. Documentation covers installation, configuration, security considerations, and development setup, indicating sustained attention to both user onboarding and contributor experience.
