Omnigram
Omnigram is a Flutter-based file reader and audiobook . It accommodates EPUB and PDF and offers audiobook functionality, supporting TTS model and other AI chat technologies for enhanced reading experiences
Last verified:
What is Omnigram?
Omnigram is an AI-native, self-hosted book library and reading service that transforms your ebook collection into an intelligent, personal reading experience. Think of it as 'Jellyfin for videos, Immich for photos, but for books.' You deploy it on your NAS or homeserver using Docker, and it manages PDF, EPUB, MOBI, AZW3, FB2, and TXT files with AI-powered features built throughout the experience.
The tool's key differentiator is its ambient AI that works invisibly in the background. When you turn to a new chapter, a context bar fades in with a recap without you pressing anything. Select a word and an AI-generated definition appears in a tooltip. The AI reading companion uses a configurable TARS personality model where you adjust proactivity, style, depth, and warmth. It includes automatic chapter recaps, inline glossary, margin notes showing cross-book connections, smart bookshelf with auto-tagging and summaries, semantic search by meaning rather than keywords, and a knowledge network that organizes notes by concept across all books.
Omnigram is designed for self-hosting enthusiasts with NAS servers (ARM or AMD64), people who have large PDF and EPUB collections they want to organize, readers who want audiobook functionality through TTS, and privacy-conscious users who need a hidden private library with biometric access. It supports multi-user management, cross-device sync for reading progress and notes, and integrates with media players like Infuse and Kodi. The app runs on Linux, Android, Windows, macOS, and iOS with native clients.
The server is built with Go 1.23 + Gin + GORM, the client uses Flutter 3.41, and it supports multiple LLM providers including OpenAI, Claude, Gemini, Ollama, and DeepSeek. TTS is handled via Kokoro-FastAPI or Edge TTS sidecar. Data portability includes exporting notes as Markdown/JSON/CSV, importing highlights from Kindle/Apple Books/Readwise, and OPDS catalog support.
Omnigram pricing
Pricing model: Freemium
Omnigram is completely free and open-source under the MIT License. There are no paid tiers, subscriptions, or premium features. The project is self-hosted so you only pay for your own hardware (NAS/server) and any external LLM provider API costs (OpenAI, Claude, Gemini, Ollama, DeepSeek) if you choose to use them. TTS uses Kokoro-FastAPI or Edge TTS which are free. Docker deployment is free. Browser extension is free. All features including AI companion, hidden library, semantic search, insights, and cross-device sync are included at no cost.
Omnigram pros
- Fully self-hosted with complete data privacy and control
- Ambient AI that works invisibly without manual activation
- Configurable AI companion personality (TARS model)
- Automatic chapter recaps that fade in when turning pages
- Inline glossary with AI-generated definitions on word selection
- Cross-book connection insights shown in page margins
- Semantic search finds content by meaning not just keywords
- Smart bookshelf auto-tags and summarizes books on import
- Hidden private library with biometric access and zero trace
- Multi-format support: EPUB, MOBI, AZW3, FB2, PDF, TXT
- Cross-platform native clients for iOS, Android, macOS, Windows, Linux
- TTS text-to-speech with customizable voices for audiobook experience
- Easy Docker one-click deployment on NAS servers
- Multi-user management with session authentication
- Reading progress and notes sync across all devices
- Knowledge network organizes notes by concept across books
- Insights page tells your reading journey story not just charts
- Export notes as Markdown, JSON, or CSV for Obsidian/Logseq
- Import highlights from Kindle, Apple Books, and Readwise
- OPDS catalog protocol for interoperability with other apps
- Encrypted local storage with platform keystore
- Integration with media players Infuse and Kodi
- Global content search across all documents on NAS
- Auto-backup files to Google Drive, OneDrive, and other clouds
Omnigram cons
- Requires self-hosting setup on NAS or homeserver
- No cloud SaaS option available
- Depends on external LLM provider for AI features
- TTS requires additional Docker sidecar container setup
- Hidden library requires biometric authentication on device
- Only 71 GitHub stars suggests relatively new project
- Documentation may be incomplete for some features
- ARM and AMD64 architecture support may limit some NAS devices
Frequently asked questions about Omnigram
What is Omnigram?
Omnigram is an AI-native, self-hosted book library and reading service. Deploy it on your NAS or homeserver using Docker, and your ebook collection becomes an intelligent, personal reading experience. Think of it as Jellyfin for videos and Immich for photos, but specifically for books. It supports EPUB, PDF, MOBI, AZW3, FB2, and TXT formats with ambient AI features built throughout.
How do I deploy Omnigram?
Omnigram-server uses Docker one-click deployment with SQLite, PostgreSQL, or MySQL support. Clone the repository, navigate to the server folder, and run 'make docker'. The server is easy to install on NAS servers with ARM or AMD64 architecture, and also works on Linux, Mac, and Windows. A demo server is available at https://omnigram-demo.lxpio.com:9443 with credentials email: admin, password: 123456.
What AI features does Omnigram have?
Omnigram has ambient AI that works invisibly: automatic chapter recaps that fade in when turning pages, inline glossary with AI definitions on word selection, margin notes showing cross-book connections, smart bookshelf with auto-tagging and summaries, AI reading companion with configurable TARS personality, semantic search by meaning, knowledge network organizing notes by concept, and reading narrative telling your journey story. You can adjust the AI's proactivity, style, depth, and warmth.
What LLM providers does Omnigram support?
Omnigram supports multiple LLM providers through langchain_dart integration: OpenAI, Claude, Gemini, Ollama, and DeepSeek. You configure your preferred provider in the environment settings. The AI features work with any OpenAI-compatible API endpoint, giving you flexibility to choose free self-hosted options like Ollama or commercial providers.
How does the TTS audiobook feature work?
Omnigram uses server-side TTS via a sidecar Docker container with OpenAI-compatible API. The default choice is Kokoro-FastAPI image which provides POST /v1/audio/speech endpoint. Edge TTS is also available as a zero-deployment fallback requiring no additional container. On-device TTS uses sherpa-onnx. You configure TTS_PROVIDER=kokoro and TTS_SIDECAR_URL in your server environment to enable audiobook functionality.
What is the hidden private library?
The hidden private library is a completely isolated second space with biometric access protection. Private books never appear in the main bookshelf, search results, or insights page - zero trace in the main interface. You get a full AI experience inside the private space including AI companion, summaries, and notes. Encrypted local storage uses platform keystore for security. This protects against others' eyes while maintaining complete functionality.
Can I sync reading progress across devices?
Yes, Omnigram syncs reading progress, notes, and highlights across all devices. The server includes a Sync Service with WebDAV support that handles progress sync, notes sync, and favorites. The Flutter client maintains local database with sqflite for offline access. Multi-user management with session authentication ensures each user's progress stays separate. You can continue reading on phone, tablet, laptop, or desktop seamlessly.
How do I export my notes and highlights?
Omnigram provides full data portability with export options: export notes and highlights as Markdown, JSON, or CSV formats. You can export your knowledge network specifically for Obsidian or Logseq integration. Full library export is available as a portable archive. OPDS catalog protocol ensures interoperability with other ebook apps. This prevents vendor lock-in and lets you use your reading data in your preferred tools.
What platforms are supported?
Omnigram is fully cross-platform. The app is available on iOS, Android, macOS, Windows, and Linux with native Flutter clients. The server runs on Linux, Mac, and Windows, and deploys on NAS servers with ARM or AMD64 architecture via Docker. Four-tab navigation includes Reading Desk, Library, Insights, and Settings. The immersive reader supports full-screen mode with customizable typography on all platforms.
Is Omnigram free?
Yes, Omnigram is completely free and open-source under the MIT License. There are no paid tiers, subscriptions, or premium features. All functionality including AI companion, hidden library, semantic search, insights, cross-device sync, and TTS audiobook are included at no cost. You only pay for your own hardware (NAS/server) and any external LLM provider API costs if you choose commercial providers over self-hosted options like Ollama.