Self-hosted ebook reader with AI-narrated audiobooks — switch seamlessly between reading and listening with word-level progress sync.

curl -sSL https://raw.githubusercontent.com/jordanful/chapter/master/install.sh | bash

Or manually:

curl -O https://raw.githubusercontent.com/jordanful/chapter/master/docker-compose.yml
docker compose up -d

Open http://localhost and create your account.

• Read & Listen — Switch between reading and audiobook mode with one tap
• Word-level sync — Your position syncs precisely between modes
• Local AI voices — High-quality TTS via Kokoro, no API costs
• 9 voices — American and British, male and female
• Library folders — Auto-import books from filesystem folders (like Plex/Calibre)
• Multi-user (coming soon) — Each user has their own library and progress
• Self-hosted — Your books, your server, your data

• Docker - Required
• CPU: 2+ cores (4+ recommended for better TTS performance)
• RAM: 2GB minimum (4GB recommended)
• Storage: 15GB minimum (30GB+ recommended)
• GPU: Optional NVIDIA GPU for 5-10x faster audiobook generation

Note: TTS generation is CPU/GPU intensive. Without GPU: ~10-15s per chapter. With GPU: ~1-2s per chapter. Audio is cached permanently.

Create a .env file next to docker-compose.yml:

# Generate with: openssl rand -base64 48
JWT_SECRET=your-secure-secret

# Audio cache size in bytes (default: 10GB)
AUDIO_CACHE_MAX_SIZE=10737418240

# Library folder limits (optional)
MAX_WATCHED_FOLDERS=20        # Max folders per user (default: 20)
SCAN_TIMEOUT=300000           # Scan timeout in ms (default: 5 min)
MAX_SCAN_DEPTH=10             # Max directory depth (default: 10)

Edit the Caddyfile section in docker-compose.yml:

yourdomain.com {
  handle /api/* {
    reverse_proxy server:3001
  }
  handle {
    reverse_proxy web:3000
  }
}

Caddy automatically provisions SSL certificates.

For faster TTS generation, uncomment the GPU section in docker-compose.yml:

kokoro:
  deploy:
    resources:
      reservations:
        devices:
          - driver: nvidia
            count: 1
            capabilities: [gpu]

Requires NVIDIA Container Toolkit.

Mount external book folders to auto-import EPUBs. Add to docker-compose.yml under server.volumes:

server:
  volumes:
    - chapter_data:/app/data
    - /path/to/your/books:/library/books:ro # Read-only mount

Then in Settings → Library, add /library/books as a watched folder. Chapter will scan and import books automatically.

Notes:

• Use container paths (e.g., /library/books), not host paths
• Mount as read-only (:ro) for safety
• Books are copied to internal storage (supports deduplication across folders)
• Multiple folders can be mounted and watched

docker compose up -d      # Start
docker compose down       # Stop
docker compose logs -f    # View logs
docker compose pull       # Update to latest

All data is in a single Docker volume. To backup:

# Database + books in one volume
docker compose cp server:/app/data ./backup

See CONTRIBUTING.md for development setup.

git clone https://github.com/jordanful/chapter.git
cd chapter
./scripts/dev-setup.sh
pnpm dev

• Self-Hosting Guide — Production deployment, backups, troubleshooting
• Architecture — System design
• Kokoro TTS — Voice configuration

GPL-2.0