Deploy and Host Karakeep on Railway

Deploy Karakeep on Railway to get a self-hosted bookmark manager with AI-powered tagging, full-text search, and browser extensions in under two minutes. Self-host Karakeep and own your data — bookmarks, notes, images, and PDFs stay on your infrastructure.

This template pre-configures three services: Karakeep (ghcr.io/karakeep-app/karakeep:release) with embedded SQLite, Meilisearch (getmeili/meilisearch:v1.41.0) for full-text search, and Browserless v2 (Railway's official railwayapp-templates/browserless-v2 build of ghcr.io/browserless/chromium:latest, fronted by Caddy for IPv6 + private-network support) for web crawling and screenshots.

Getting Started with Karakeep on Railway

Open the generated public URL after deployment completes. Create your first account — this becomes the admin user. Navigate to Settings to add your OpenAI API key for automatic tagging.

Install the browser extension from Chrome Web Store or Firefox Add-ons, enter your server URL and API key in extension settings. Start saving links, images, or notes — Karakeep automatically crawls URLs, captures screenshots, and applies AI tags. Set DISABLE_SIGNUPS=true in Railway after creating your account to block further registration.

About Hosting Karakeep on Railway

Karakeep (formerly Hoarder) is an open-source, self-hostable bookmark-everything application built for developers and researchers who want intelligent organization without cloud dependencies.

Key features:

• Bookmark links, notes, images, and PDFs in one place
• AI-powered automatic tagging via OpenAI or Ollama
• Full-text search powered by Meilisearch
• Browser extensions for Chrome and Firefox
• Mobile apps for iOS and Android
• RSS feed import, OAuth/OIDC, and REST API
• Rule-based auto-organization into lists

Why Deploy Karakeep on Railway

Railway eliminates the complexity of orchestrating three services, volumes, and private networking.

• One-click deploy with pre-configured Meilisearch and headless browser
• Persistent volumes for SQLite database and search indexes
• Private networking — internal APIs never publicly exposed
• Automatic HTTPS with zero configuration
• Pay only for actual resource usage

Common Use Cases for Self-Hosted Karakeep

• Developer knowledge base — save Stack Overflow solutions and docs with AI-tagged categories by language and framework
• Research library — collect papers and articles with full-text search across archived content
• Design inspiration — bookmark screenshots with OCR-extracted text becoming searchable
• Team shared bookmarks — collaborative link collection with auto-organization rules

Dependencies for Karakeep on Railway

• Karakeep — ghcr.io/karakeep-app/karakeep:release — web UI, API, and background workers
• Meilisearch — getmeili/meilisearch:v1.41.0 — full-text search engine
• BrowserlessV2 — railwayapp-templates/browserless-v2 (GitHub repo build of ghcr.io/browserless/chromium:latest + Caddy reverse proxy) — headless Chromium with WebSocket Playwright endpoint and token auth

Environment Variables Reference for Karakeep

Deployment Dependencies

• GitHub (Karakeep): karakeep-app/karakeep
• GitHub (Browserless template): railwayapp-templates/browserless-v2
• Docker: ghcr.io/karakeep-app/karakeep
• Docs: docs.karakeep.app

Hardware Requirements for Self-Hosting Karakeep

Meilisearch needs ~500 MB RAM for indexing and Browserless v2 uses 300-500 MB per crawl session.

Self-Hosting Karakeep with Docker Compose

Create a docker-compose.yml:

services:
  karakeep:
    image: ghcr.io/karakeep-app/karakeep:release
    ports:
      - "3000:3000"
    volumes:
      - karakeep_data:/data
    environment:
      - DATA_DIR=/data
      - NEXTAUTH_URL=http://localhost:3000
      - NEXTAUTH_SECRET=change-me-to-random-string
      - MEILI_ADDR=http://meilisearch:7700
      - MEILI_MASTER_KEY=your-master-key
      - BROWSER_WEBSOCKET_URL=ws://browserless:3000/chromium/playwright?token=your-browser-token
      - BROWSER_CONNECT_ONDEMAND=true
    depends_on:
      - meilisearch
      - browserless
  meilisearch:
    image: getmeili/meilisearch:v1.41.0
    volumes:
      - meili_data:/meili_data
    environment:
      - MEILI_MASTER_KEY=your-master-key
      - MEILI_NO_ANALYTICS=true
  browserless:
    image: ghcr.io/browserless/chromium:latest
    environment:
      - TOKEN=your-browser-token
      - TIMEOUT=120000
volumes:
  karakeep_data:
  meili_data:

Then run:

docker compose up -d

How Much Does Karakeep Cost to Self-Host?

Karakeep is open-source under the AGPL license — free forever. On Railway, expect $5-10/month on the Hobby plan. Karakeep Cloud costs $4/month with 50 GB storage, while self-hosting gives unlimited volume-based storage and no feature restrictions.

Karakeep vs Raindrop.io vs Linkwarden

Karakeep uses lightweight SQLite instead of PostgreSQL, supports fully local AI via Ollama, and provides native mobile apps — the easiest self-hosted option to deploy.

FAQ

What is Karakeep and why self-host it on Railway? Karakeep is an open-source bookmark manager for links, notes, images, and PDFs with AI-powered auto-tagging and full-text search. Self-hosting on Railway gives you full data ownership and a pre-configured multi-service architecture that deploys in one click.

What does this Railway template deploy for Karakeep? Three services: the main Karakeep application (web UI, API, workers with embedded SQLite), Meilisearch for full-text search, and Browserless v2 (Railway's official railwayapp-templates/browserless-v2 build) for web crawling via the Playwright WebSocket protocol. Two persistent volumes store the database/assets and search indexes.

Why does Karakeep need Meilisearch and a headless browser? Meilisearch provides sub-50ms full-text search across all bookmarked content. Browserless v2 ships a current Chromium build with Playwright support so Karakeep's crawler can render JavaScript-heavy pages, capture screenshots, archive HTML, and extract metadata for the search index. The connection is token-authenticated and stays on Railway's private network.

How do I enable AI tagging in self-hosted Karakeep on Railway? Add your OpenAI API key as the OPENAI_API_KEY environment variable in Railway. For fully local AI, deploy Ollama and set OLLAMA_BASE_URL. AI tagging applies automatically to new bookmarks.

Can I migrate from Pocket or Raindrop.io to Karakeep? Yes. Karakeep imports from Pocket (HTML export), Raindrop.io (CSV), browser bookmarks (HTML), and Omnivore via Settings > Import.

How do I connect the Karakeep browser extension after deploying on Railway? Install from Chrome Web Store or Firefox Add-ons, enter your Railway public URL, then generate an API key from Settings > API Keys and paste it into the extension.

Is Karakeep suitable for team use with multiple users? Yes. Keep DISABLE_SIGNUPS=false for open registration or configure OAuth/OIDC for SSO. Shared lists enable team collaboration on curated link collections.