Karakeep

Karakeep lets users store links, notes, images, PDFs and other web content in a self‑hosted collection. It automatically retrieves titles, descriptions and preview images for saved URLs, archives full pages to guard against link rot, and can extract text from images via OCR. Content is indexed for full‑text search and can be organized into lists, with bulk actions and importers for browsers and other bookmark services.

The application adds AI‑driven tagging and summarization, supporting both cloud models like OpenAI and local models through Ollama. A rule‑based engine enables custom automation, while RSS feed ingestion and video archiving extend its hoarding capabilities. Collaboration is possible through shared lists, SSO, and a REST API with webhooks for integration.

Native extensions for Chrome and Firefox, plus iOS and Android apps, provide quick bookmarking from any device. The stack uses NextJS, Drizzle, NextAuth, tRPC, Puppeteer, Meilisearch and is distributed under the AGPL‑3.0 license, offering a free, non‑subscription, open‑source solution that can be deployed via Docker.