Getting Started

Getting Started

System Requirements

Requirement	Minimum	Recommended
OS	Windows 10, macOS 12, Linux (glibc 2.31+)	Latest stable
RAM	4 GB	8 GB+
Disk	500 MB (app) + dataset space	SSD recommended
GPU	—	NVIDIA with CUDA or Apple Silicon (MPS)
FFmpeg	Required for video annotation	Bundled in release builds
Chromium	Optional (browser automation)	Chrome, Brave, Edge, or Chromium

Prerequisites for Development

1. Node.js >= 18
2. Rust >= 1.89
3. Tauri 2 system dependencies — follow the official guide for your platform:
   • Tauri 2 Prerequisites

Linux-specific packages

# Ubuntu / Debian
sudo apt install libwebkit2gtk-4.1-dev libappindicator3-dev librsvg2-dev patchelf pkg-config

# Fedora
sudo dnf install webkit2gtk4.1-devel libappindicator-gtk3-devel librsvg2-devel

Windows-specific

• Visual Studio Build Tools (or full Visual Studio) with MSVC
• WebView2 (pre-installed on Windows 11, available for Windows 10)

macOS-specific

xcode-select --install

Installation

# Clone the repository
git clone https://github.com/tecmedhub/annotix.git
cd annotix

# Install frontend dependencies
npm install

Development

# Full development mode (React hot-reload + Rust auto-rebuild)
npm run tauri:dev

# Frontend only (Vite dev server on port 5173)
npm run dev

When running tauri:dev, the Rust backend compiles on first launch (this takes a few minutes). Subsequent rebuilds are incremental and much faster.

Production Build

npm run tauri:build

Build outputs by platform:

Platform	Output
Linux	AppImage + raw binary in src-tauri/target/release/bundle/
Windows	NSIS installer (.exe) + MSI in src-tauri/target/release/bundle/
macOS	.app bundle + DMG in src-tauri/target/release/bundle/

Available Scripts

Script	Description
npm run dev	Frontend only (Vite dev server)
npm run build	TypeScript check + Vite production build
npm run tauri:dev	Full development with Rust backend
npm run tauri:build	Production build with platform-specific installers
npm run lint	ESLint with zero-warnings policy
npm run preview	Preview the built frontend

First Run

1. Launch Annotix.
2. The setup screen appears automatically on first run, asking you to choose a directory where projects will be stored.
3. Choose a directory (e.g. ~/annotix-projects). This is saved to ~/.local/share/annotix/config.json.
4. You're ready to create your first project.

Creating Your First Project

1. Click New Project on the project list screen.
2. Enter a name and select a project type (e.g. bbox for object detection).
3. Define your classes (e.g. "car", "person") with colors.
4. Upload images via drag-and-drop or the upload button.
5. Click an image to open the annotation canvas and start drawing.

Project Types

Images

Type	Use Case
bbox	Object detection with rectangular bounding boxes
obb	Oriented/rotated bounding box detection
polygon	Semantic segmentation with polygons
mask	Semantic segmentation with freehand painting
instance-segmentation	Instance segmentation (mask + polygon)
keypoints	Pose estimation with skeleton presets
landmarks	Named reference points
classification	Single-label image classification
multi-label-classification	Multi-label image classification

Time Series

Type	Use Case
timeseries-classification	Classify entire series
timeseries-forecasting	Predict future values
anomaly-detection	Detect anomalous points/ranges
timeseries-segmentation	Segment temporal regions
pattern-recognition	Find recurring patterns
event-detection	Mark discrete events
timeseries-regression	Continuous value prediction
clustering	Group similar series
imputation	Fill missing values

Other

Type	Use Case
tabular	Classical ML on structured data

Application Routes

Route	View
/	Project list
/projects/:id	Image gallery + class management
/projects/:id/images/:imageId	Annotation canvas
/projects/:id/timeseries/:tsId	Time series annotation
/projects/:id/videos/:videoId	Video annotation with timeline
/settings	Application settings

Storage Layout

All data is stored as JSON files and raw assets on disk. No database.

~/.local/share/annotix/
  config.json                          -> global config (projects_dir path)

{projects_dir}/{uuid}/
  project.json                         -> full project (metadata, classes, annotations)
  images/{uuid}_{filename}             -> original images
  thumbnails/{id}.jpg                  -> generated thumbnails
  videos/{uuid}_{filename}             -> video files
  models/{uuid}_{filename}             -> inference models

Next Steps

• Annotation Tools — Learn all 7 annotation tools in detail.
• Integrated ML Training — Set up your Python environment and train models.
• P2P Collaboration — Collaborate with your team in real-time.