karakeep-app/karakeep

A self-hostable bookmark-everything app (links, notes and images) with AI-based automatic tagging and full text search

What it solves

Karakeep is a self-hostable bookmarking and knowledge management app designed for "data hoarders" who need a central place to store links, notes, images, and PDFs. It addresses the problem of link rot and disorganized content by providing full-page archival and automatic organization tools.

How it works

The app uses a combination of web crawling (Puppeteer) and AI to manage content. It automatically fetches titles, descriptions, and images from links. It employs LLMs (via OpenAI or local models through Ollama) to automatically tag and summarize content, and uses Meilisearch for full-text and semantic search. It also integrates with tools like monolith for full-page archival and yt-dlp for video archiving.

Who it’s for

It is built for users who prefer self-hosting their data and want a sophisticated "read-it-later" system that can automatically organize and summarize their saved content.

Highlights

  • AI-Powered Organization: Automatic tagging and summarization using LLMs, with support for local models via Ollama.
  • ** Cortana/Agent Friendly**: Includes a CLI and official skills for LLM agents like OpenClaw and Hermes.
  • Comprehensive Archiving: Full-page archival to prevent link rot and automatic video archiving.
  • Crosspoint Accessibility: Available via web, Chrome/Firefox/Safari extensions, and iOS/Android apps.
  • Semantic Search: Supports both full-text and semantic search across all stored content.
  • OCR: Extracts text from images for better searchability.

Related

  • Project
    jihe520/mindpocket

    A serverless, AI-powered personal bookmarking system that uses RAG for automatic summarization and tagging across web, mobile, and browser extensions.

  • Project
    sidinsearch/superbrain

    A self-hosted AI-powered second brain for Android that archives, summarizes, and organizes content from Instagram, YouTube, and the web.

  • Project
    WangQrkkk/PaperQuay

    An open-source, local-first AI paper workspace that integrates PDF reading, translation, structured overviews, and rich note-taking for researchers.

  • Project
    Future-Scholars/paperlib

    An open-source academic paper management tool that simplifies metadata scraping and organization, featuring LLM-powered summarization and semantic search.

  • Project
    AeternaLabsHQ/pullmd

    PullMD is a self‑hosted service that converts URLs, documents, images, audio, and YouTube videos into clean Markdown with YAML front‑matter. It runs via Docker, uses optional side‑cars for HTML extraction, headless rendering, and document conversion, and offers an API, PWA UI, and share‑link caching—aimed at AI agents and developers needing token‑efficient content.