Skip to content
BerryUIKIPublic

About

A comprehensive image metadata indexer and viewer designed specifically for AI-generated images. Organize, search, and manage your AI-generated media collection with powerful metadata tools.

Resources

Contributing

Stars

98 stars

Watchers

7 watching

Forks

Latest commit

 

History

728 Commits

Folders and files

Repository files navigation

🍇 Omera

The open-source asset manager & prompt studio for AI-generated images.

Organize, search, compare, and export tens of thousands of AIGC artworks — all from a blazing-fast local desktop app.


English | 简体中文 | 繁體中文 | 日本語


GitHub Stars   GitHub Forks   GitHub Issues


Release Downloads License Website

Tauri 2 Rust Vue 3 Platform


Omera — 3-Pane Studio Workspace

Illustrative UI mockup · v0.5.0


🚀 Quick Start

  1. Download — Grab the latest release for your platform.
  2. Set Up — Choose the application language and database directory. Review any discovered Berry library before importing; restart to activate a verified snapshot. Cleanup is optional and requires separate confirmation.
  3. Add a Folder — Point Omera at your WebUI / ComfyUI / NovelAI output directory. Choose External Link (zero-copy) or AIGC Pipeline (auto-harvest) mode.
  4. Browse & Create — Your library is instantly searchable. Rate, tag, compare, and export with full metadata preserved.

🆕 What's New in v0.5.0

Flux.1 / SD3.5 parsing, direct Civitai integration, and signed 0.5.0 installers — Read the full changelog →

  • 🧬 Flux.1 / SD3.5 parameter parsing — Guidance, Distilled CFG Scale, Shift, schedule type and dual CLIP-L / T5-XXL prompts from A1111/Forge and ComfyUI metadata.
  • 🔎 Civitai lookups — Reverse-lookup checkpoint and LoRA hashes, cache preview cards locally, flag newer published versions, and merge trained trigger words into your LoRA catalog.
  • 🛡️ Signed release — v0.5.0 ships signed Windows, macOS and Linux installers with matching .minisig assets.
  • 🎨 Clearer generation details — The Inspector surfaces the new Flux/SD3.5 fields and dual text-encoder prompts in all seven languages.

✨ Key Features

🎨 All-in-One 3-Pane Studio Workspace

  • Frameless Window with Native Quality — Custom title bar with integrated desktop menu bar (File, Edit, View, Tools, Help), drag region, and window controls.
  • Left Navigation Sidebar — Quick filters (All Images, Favorites, Sensitive 18+), hierarchical folder tree with real-time scan indicators, color-coded tags, and smart albums.
  • Center Canvas — Grid · Waterfall · Table — Ultra-fast virtual scrolling rendering tens of thousands of images. Masonry Waterfall view, smooth thumbnail zoom slider (130 px – 360 px), and one-click ⊞ / ▦ / ☰ view switcher.
  • Stable Responsive Density — Fixed card widths during resize; columns are added or removed instead of stretching images.
  • Right Property Inspector — Large preview card, star ratings (0–5), favorite toggle, tokenized prompt chips with one-click copy, LoRA tags, generation parameters, and collapsible raw workflow JSON.
  • Immersive Lightbox (Quick Look) — Full-screen viewer (Space / Enter) with smooth mouse-wheel zoom, pan, and keyboard navigation.

🔍 Lossless AIGC Metadata Parsers

Automatically extracts and indexes Prompt, Negative Prompt, Model, Hash, Sampler, Steps, CFG, Seed, Dimensions, and full Workflow JSON:

Platform Source
AUTOMATIC1111 / SD.Next PNG tEXt/iTXt parameters, WebP EXIF
ComfyUI Full prompt & workflow graph JSON, LoRA Loader detection
NovelAI Comment & Description signature decoding
Fooocus / Fooocus-MRE Parameter parsing & model resolution
InvokeAI & EasyDiffusion Embedded metadata & JSON sidecars
Flux.1 / SD3.5 A1111/Forge guidance fields, ComfyUI dual-encoder prompts
Sidecar Files .txt companion metadata

Supported formats: PNG · JPG/JPEG · WebP · AVIF · MP4 · WebM

🎬 Video & Animation Support

  • AIGC video metadata — MP4/WebM containers with ComfyUI AnimateDiff, Wan2.1, HunyuanVideo, CogVideoX, LTX-Video and SVD workflows.
  • Hover scrubbing — Scrub through video frames directly on Grid and Waterfall cards.
  • Fullscreen player — Loop, 0.25x–2x speed, frame-by-frame stepping (. / ,), mute, and the shared Inspector.
  • Video facets — Filter by duration range, frame rate (FPS) and media type; resolution, container-codec and motion-score filters are planned.

🗃️ Multi-Mode Ingestion & Smart Stacking

  • External Link — Reference in-place, zero-copy.
  • Managed Vault — Dedicated app-managed storage.
  • AIGC Pipeline — Automatic background harvest from WebUI/ComfyUI output folders with debounced ingestion.
  • Intelligent Burst Stacking — Auto-groups sequential generations with similar prompts.
  • Interactive Stack Cards — Poker-deck visual cards with badge counts, inline expand/collapse, cover selection (Alt+S), side-by-side compare (C), and safe transactional stack merging.

🧠 AI Intelligence

  • Local CLIP / SigLIP Semantic Search — Search images using natural language via local ONNX models — no cloud required.
  • WD14 / Danbooru Auto-Tagging — Automated anime & realistic tag extraction with confidence thresholds.
  • Visual Similarity Search — Find compositionally similar images from any starting point.
  • LoRA Trigger Word Library — Scan LoRA directories, merge Civitai trained words, descriptions and previews, and one-click inject trigger words.
  • Prompt Keyword Insights — Frequency analysis of prompt tokens with average rating correlations.
  • Checkpoint Model Manager — Direct Civitai lookups by AutoV2/SHA256, cached preview cards, version update badges, and one-click model filtering.
Omera Model Manager with Civitai lookups, cached previews and update badges

Illustrative UI mockup · Civitai lookups, cached previews and update badges

☁️ Cloud Sync & Export

  • S3 / WebDAV Snapshot Backup — Hot VACUUM INTO snapshots to AWS S3, Cloudflare R2, MinIO, Backblaze B2, and WebDAV servers with schema verification and rollback safety.
  • Incremental Remote Media Mirroring (Upload-Only) — One-way library media upload mirroring to S3, WebDAV, and network paths with file-size and SHA-256 change detection, token-bucket bandwidth throttling, dry-run simulation, and cancelable multi-threading. Remote-to-local pull, ETag delta reconciliation, and bidirectional sync planned; this release remains upload-only.
  • Batch Export & Transcoding — Multi-threaded WebP/JPEG/PNG conversion with 4-tier privacy metadata stripping and customizable filename templates.
  • HTML Showcase Generator — Standalone zero-dependency index.html with responsive dark gallery, fullscreen lightbox, prompt inspector, and instant keyword filter.

👥 Multi-Database Team Studio (Planned)

  • Current Runtime: SQLite — Fast local database with WAL mode for single-user and network file system scenarios.
  • Storage Engine Abstraction — StorageEngine trait and dialect system designed for multi-backend support. MySQL 8.0+ and PostgreSQL 14+ backends are experimental/planned.
  • Cross-Platform Storage Roots — Path normalization for multi-user collaboration across Windows, macOS, and Linux.
  • SQL Export & Connection Testing — Export SQLite data for migration; test remote connection strings (implementation of live query backends is planned).

🌐 Internationalization & Auto-Update

  • 7 Languages — English · 简体中文 · 繁體中文 · 日本語 · Deutsch · Français · Español
  • Auto OS Language Detection — Follows your system language by default.
  • Built-in Updater — Check for updates directly from Help > Check for Updates… with release notes and one-click download.

⚡ Performance

Omera is built for large libraries. The gallery renders from the indexed SQLite database before any filesystem reconciliation begins — no blocking startup scan.

Metric 1 k images 10 k images 50 k images
Time to first usable gallery (warm DB) 4.48 ms 23.65 ms 143.55 ms
Main-thread long tasks (> 50 ms) during rapid scrollbar drag 0 0 0
Thumbnail manifest adoption — — 3.26 s (15,300 files/sec)

Additional safeguards:

  • O(1) Keyset Cursor Pagination — < 1 ms per page at 500 k+ assets (84× faster than OFFSET).
  • Frame-Coalesced Scrolling — At most 1 reactive update per requestAnimationFrame.
  • Per-Column Binary Search — Waterfall visibility computed via binary search, not full-library scan.
  • Demand-Driven Thumbnails — Visible images decoded first; look-ahead starts only after scrolling settles.
  • Directory Fingerprint Gating — Parent mtime check delivers 2.5× local / 5.6× network speedup on re-scans.

📖 See Performance Architecture & Benchmarks for full details.


🏗️ Architecture

Omera architecture: Vue 3 frontend over Tauri IPC, the omera app shell, six omera-* Rust crates, and external SQLite, asset, watcher, Civitai, update and cloud services

Tech Stack: Tauri 2 · Rust (2021 edition) · Vue 3.5 · TypeScript 5.6 · Vite 6 · SQLite (WAL) · Rayon · ONNX Runtime


⌨️ Keyboard Shortcuts

Shortcut Action Shortcut Action
Space / Enter Open / Close Lightbox 0 – 5 Set Star Rating (0 = Clear)
F Toggle Favorite B Toggle Sidebar
I Toggle Inspector / or Ctrl+F Focus Search Bar
Ctrl+A Select loaded items Esc Clear Selection / Close Modal
Ctrl+O Folder Wizard Ctrl+, Settings
Ctrl+G Merge Stacks Ctrl+Shift+G Unstack
Alt+S Set Stack Cover C Side-by-Side Compare
Delete Move to Trash ? Shortcuts Guide

📥 Downloads

Get the latest pre-built binaries from GitHub Releases:

Platform Architecture Format File Name
Windows x86_64 NSIS Installer Omera_<version>_x64-setup.exe
Windows x86_64 MSI Omera_<version>_x64_en-US.msi
macOS Apple Silicon (arm64) DMG Omera_<version>_aarch64.dmg
Linux x86_64 AppImage Omera_<version>_amd64.AppImage
Linux x86_64 Debian Omera_<version>_amd64.deb
Linux x86_64 RPM Omera-<version>-1.x86_64.rpm

The current release matrix covers Windows x64, macOS Apple Silicon and Linux x64. Intel macOS and a portable ZIP are not published.


🛠️ Building from Source

Prerequisites

  • Node.js v18+ & pnpm — npm install -g pnpm
  • Rust 1.75+ — Install via rustup.rs
  • Platform Toolchain — MSVC Build Tools (Windows), Xcode CLI (macOS), libwebkit2gtk-4.1 (Linux)

Build

# Clone
git clone https://github.com/BerryUIKI/Omera.git
cd Omera

# Install dependencies
pnpm install

# Development (hot-reload)
pnpm dev:desktop

# Production build
pnpm tauri build

Output: src-tauri/target/release/bundle/


🗺️ Roadmap

Omera roadmap status at v0.5.0: thirteen milestones complete, video and model-ecosystem milestones partially complete, and milestones 17–19 planned

Milestone 15 (video) and Milestone 16 (Flux.1 / SD3.5 + Civitai) are partially complete. The remaining items are resolution/codec/motion-score facets and the ComfyUI workflow visualizer. Milestone 20 (image compression) is delivered; Milestones 17–19 cover real-time ComfyUI workflows, the LAN web companion and color analytics.

📖 See docs/ROADMAP.md for the full checklist.


🤝 Contributing

We welcome bug reports, feature requests, new metadata parsers, and translation contributions!

  1. Fork the repo and branch from dev (not main).
  2. Read the Contributing Guide and Agent Guidelines.
  3. Ensure all checks pass before opening a PR:
pnpm run build          # Frontend build
pnpm run test:stack     # Behavior tests
cargo clippy --workspace -- -D warnings
cargo test --workspace  # Rust tests

📄 License

This project is licensed under the Apache-2.0 License.

Copyright © 2026 BerryUIKI.


🙏 Acknowledgments

Omera is built on the shoulders of these excellent open-source projects:


⬆ Back to Top

If you find Omera useful, please consider giving it a ⭐ — it helps others discover the project!

Star History Chart

Made with ❤️ by the Omera community

About

A comprehensive image metadata indexer and viewer designed specifically for AI-generated images. Organize, search, and manage your AI-generated media collection with powerful metadata tools.

Resources

Contributing

Stars

98 stars

Watchers

7 watching

Forks

Releases

Packages

Contributors

Languages