Show HN: Open-source AI bookmark manager for busy people

Cache, an open-source AI bookmark manager, unifies bookmarks across platforms into a single actionable library with AI-assisted organization, search, and automations. The tool, built with Next.js, Bun, PostgreSQL, and Google Gemini, is available as a cloud-hosted service at www.cachd.app or for self-hosting with zero telemetry.

Unify your bookmarks across platforms into a single actionable library. Features features • Why why • Tech Stack tech-stack • MCP cache-mcp • Roadmap roadmap • Contributing contributing • License license Bookmarking is broken. When you hit "save" on a tweet, a video, or a post, you are making a deliberate decision that this is worth remembering . But that intent is immediately lost. It vanishes into a list you never revisit, scattered across a dozen platforms with no connection to your actual workflow or goals. The feeds are designed to keep you scrolling, not to help you resurface what you need. Existing tools treat the "save" action as an afterthought, a dead end rather than a starting point. Cache exists because that signal is too valuable to waste. It treats the act of saving as a first-class event and builds the entire experience around turning that intent into action. It does not replace your platforms; it respects the intent behind why you use them and gives it a destination. Unify your bookmarks — Integrate Cache into your day-to-day with first-class support for bookmarks from Browser bookmarks, Instagram Saved, TikTok Favorites, YouTube Watch Later, X/Twitter bookmarks, GitHub Stars, Pinterest, Google Photos, MCP, and more, all in one place. Unlike other tools that cap saves, Cache has no limits. Smart collections — Automatically organizes entries into your collections with AI-assisted relevance ranking. Cache even learns your preferences over time. Overviews — See a 1-line summary above every collection. As new entries are added, it updates instantly. And if you want to see more detail, just hit expand. AI-assisted search — Ask the Cache AI agent and search across all your saved content. Automations — Create custom agents to do anything. Generate daily digests, summaries, weekly reminders, and much more. Note-taking — First-party note-taking support alongside bookmarks. Collaboration — Share a live view of any collection with anyone, even if they don't use Cache. Browser extension — Capture and sync saved content from anywhere on the web. Export & integrate — Pipe results into other tools you already use. Simple and low maintenance — Cache is designed to be simple, low-maintenance, and always portable. Cloud-hosted: www.cachd.app https://www.cachd.app You can self-host Cache for total control over your data and design. Cache has zero telemetry by default. Bun https://bun.sh/ v1.3.14+ Node.js https://nodejs.org/ 24.x- PostgreSQL 12+ local or remote - A Google Gemini API key for AI features Clone the repository git clone https://github.com/rortan134/cache-app.git cd cache Install dependencies bun install Set up environment cp .env.example .env Edit .env with your database URL and API keys Set up the database bun run db-deploy Start the development server bun run dev Open http://localhost:3000 http://localhost:3000 . See the environment variables reference /rortan134/cache-app/blob/main/.env.example for the full list. Next.js · Bun · PostgreSQL · Prisma ORM · Better Auth · Tailwind — and more | Category | Technology | |---|---| Framework | | UI React https://react.dev/ , Base UI https://base-ui.com/ , Tailwind CSS https://tailwindcss.com/ Rich Text Lexical https://lexical.dev/ , Streamdown https://github.com/vercel/streamdown Database Auth Better Auth https://better-auth.com/ Validation Zod https://zod.dev/ , @t3-oss/env-nextjs https://env.t3.gg/ AI/LLM AI SDK https://sdk.vercel.ai/ , AI Gateway https://vercel.com/ai-gateway , Gemini https://ai.google.dev/ Agentic Web Search Tavily https://tavily.com/ Data Fetching SWR https://swr.vercel.app/ , nuqs https://nuqs.vercel.app/ i18n gt-next https://generaltranslation.com Subscriptions Stripe https://stripe.com/ Durable Execution Workflow SDK https://workflow-sdk.dev/ Security Cloud-only Arcjet https://arcjet.com/ WAF, rate limiting, PII redaction Linting Ultracite https://ultracite.dev/ Biome Date Handling Day.js https://day.js.org/ , chrono-node https://github.com/wanasit/chrono Cache exposes an MCP https://modelcontextprotocol.io/ server so AI agents like Claude, Cursor, and others can read and write your library directly. Search bookmarks, save new items, list collections, and more. Endpoint: https://www.cachd.app/mcp llms.txt https://www.cachd.app/llms.txt — agent context and tool reference- Generate a setup prompt with your Bearer token from the app Integrations → MCP Remind me — Set up unique reminders when saving or browsing on items to come back to later. Comments — Add and view threaded comments on entries. Inbox view — Triage view for reviewing entries. Notes improvements — Richer editing experience, advanced formatting. Raycast integration — Capture and search Cache from Raycast. Substack integration — Import and save Substack posts and newsletters. We welcome contributions Please see our Contributing Guide /rortan134/cache-app/blob/main/CONTRIBUTING.md for details. Open an issue https://github.com/rortan134/cache-app/issues?q=sort%3Aupdated-desc+is%3Aissue+state%3Aopen+ if you believe you've encountered a bug. This project follows the Contributor Covenant /rortan134/cache-app/blob/main/CODE OF CONDUCT.md code of conduct. This project is licensed under the Apache License 2.0 - see the LICENSE file /rortan134/cache-app/blob/main/LICENSE for details.