SnapOtter/apps/docs/public/llms-full.txt

# Stirling Image — Complete Documentation

Self-hosted, open-source image processing platform with 30+ tools. Runs in a single Docker container. Resize, compress, convert, remove backgrounds, upscale, run OCR, and more — without sending images to external services.

- GitHub: https://github.com/stirling-image/stirling-image
- Docs: https://stirling-image.github.io/stirling-image/
- License: MIT

---

# Getting Started

## Run with Docker

The fastest way to get Stirling Image running:

```bash
docker run -d \
  --name stirling-image \
  -p 1349:1349 \
  -v stirling-data:/data \
  stirlingimage/stirling-image:latest
```

Open `http://localhost:1349` in your browser. Log in with `admin` / `admin`.

## Run with Docker Compose

Create a `docker-compose.yml`:

```yaml
services:
  stirling-image:
    image: stirlingimage/stirling-image:latest
    container_name: stirling-image
    ports:
      - "1349:1349"
    volumes:
      - stirling-data:/data
      - stirling-workspace:/tmp/workspace
    environment:
      - AUTH_ENABLED=true
      - DEFAULT_USERNAME=admin
      - DEFAULT_PASSWORD=admin
    restart: unless-stopped

volumes:
  stirling-data:
  stirling-workspace:
```

```bash
docker compose up -d
```

## Build from source

Requirements: Node.js 22+, pnpm 9+, Python 3.10+

```bash
git clone https://github.com/stirling-image/stirling-image.git
cd stirling-image
pnpm install
pnpm dev
```

This starts both the API server and the React frontend at `http://localhost:1349`.

---

# Architecture

Stirling Image is a monorepo managed with pnpm workspaces and Turborepo. Everything ships as a single Docker container.

## Project structure

```
Stirling-Image/
├── apps/
│   ├── api/          # Fastify backend
│   ├── web/          # React + Vite frontend
│   └── docs/         # VitePress documentation site
├── packages/
│   ├── image-engine/ # Sharp-based image operations
│   ├── ai/           # Python AI model bridge
│   └── shared/       # Types, constants, i18n
└── docker/           # Dockerfile and Compose config
```

## Packages

### @stirling-image/image-engine

Core image processing library built on Sharp. Handles all non-AI operations: resize, crop, rotate, flip, convert, compress, strip metadata, and color adjustments. No network dependencies, runs entirely in-process.

### @stirling-image/ai

Bridge layer that calls Python scripts via child processes. Each AI capability has a TypeScript wrapper that spawns a Python subprocess, passes image data through the filesystem, and returns the result.

Supported operations:
- Background removal — BiRefNet-Lite model via rembg
- Upscaling — RealESRGAN
- OCR — PaddleOCR
- Face detection/blurring — MediaPipe
- Object erasing (inpainting) — LaMa Cleaner

Python scripts live in `packages/ai/python/`. The Docker image pre-downloads all model weights during the build so the container works offline.

### @stirling-image/shared

Shared TypeScript types, constants (like APP_VERSION and tool definitions), and i18n translation strings used by both frontend and backend.

## How a request flows

1. The user picks a tool in the web UI and uploads an image.
2. The frontend sends a multipart POST to `/api/v1/tools/:toolId` with the file and settings.
3. The API route validates the input with Zod, auto-orients the image based on EXIF metadata, then calls the appropriate package function.
4. For AI tools, the TypeScript bridge spawns a Python subprocess, waits for it to finish, and reads the output file.
5. The API returns a `jobId` and `downloadUrl`. The frontend can poll `/api/v1/jobs/:jobId/progress` via SSE for real time status.
6. The user downloads the processed image from `/api/v1/download/:jobId/:filename`.

---

# Configuration

All configuration is done through environment variables. Every variable has a sensible default.

## Server

| Variable | Default | Description |
|---|---|---|
| PORT | 1349 | Port the server listens on |
| RATE_LIMIT_PER_MIN | 100 | Maximum requests per minute per IP |

## Authentication

| Variable | Default | Description |
|---|---|---|
| AUTH_ENABLED | false | Set to true to require login. Docker image defaults to true. |
| DEFAULT_USERNAME | admin | Username for the initial admin account. Only used on first run. |
| DEFAULT_PASSWORD | admin | Password for the initial admin account. Change after first login. |

## Storage

| Variable | Default | Description |
|---|---|---|
| STORAGE_MODE | local | local or s3. Only local is currently implemented. |
| DB_PATH | ./data/stirling.db | Path to the SQLite database file |
| WORKSPACE_PATH | ./tmp/workspace | Directory for temporary files during processing |
| FILES_STORAGE_PATH | ./data/files | Directory for persistent user files |

## Processing limits

| Variable | Default | Description |
|---|---|---|
| MAX_UPLOAD_SIZE_MB | 100 | Maximum file size per upload in megabytes |
| MAX_BATCH_SIZE | 200 | Maximum number of files in a single batch request |
| CONCURRENT_JOBS | 3 | Number of batch jobs that run in parallel |
| MAX_MEGAPIXELS | 100 | Maximum image resolution allowed |

## Cleanup

| Variable | Default | Description |
|---|---|---|
| FILE_MAX_AGE_HOURS | 24 | How long temporary files are kept before deletion |
| CLEANUP_INTERVAL_MINUTES | 30 | How often the cleanup job runs |

## Appearance

| Variable | Default | Description |
|---|---|---|
| APP_NAME | Stirling Image | Display name shown in the UI |
| DEFAULT_THEME | light | Default theme (light or dark) |
| DEFAULT_LOCALE | en | Default interface language |

## Volumes

- `/data` — Persistent storage for SQLite database and user files. Mount to keep data across restarts.
- `/tmp/workspace` — Temporary storage for images being processed. Optional but prevents container layer growth.

---

# Database

SQLite with Drizzle ORM. Database file at DB_PATH (default: ./data/stirling.db).

## Tables

### users
| Column | Type | Notes |
|---|---|---|
| id | integer | Primary key, auto-increment |
| username | text | Unique, required |
| passwordHash | text | bcrypt hash |
| role | text | admin or user |
| mustChangePassword | integer | Boolean flag for forced password reset |
| createdAt | text | ISO timestamp |

### sessions
| Column | Type | Notes |
|---|---|---|
| id | text | Primary key (session token) |
| userId | integer | Foreign key to users.id |
| expiresAt | text | ISO timestamp |

### api_keys
| Column | Type | Notes |
|---|---|---|
| id | integer | Primary key |
| userId | integer | Foreign key to users.id |
| keyHash | text | SHA-256 hash of the key |
| name | text | User-provided label |
| createdAt | text | ISO timestamp |
| lastUsedAt | text | Updated on each request |

Keys are prefixed with `si_` followed by 96 hex characters.

### pipelines
| Column | Type | Notes |
|---|---|---|
| id | integer | Primary key |
| name | text | Pipeline name |
| description | text | Optional |
| steps | text | JSON array of { toolId, settings } |
| createdAt | text | ISO timestamp |

### settings
| Column | Type | Notes |
|---|---|---|
| key | text | Primary key |
| value | text | Setting value |

---

# Deployment

## Docker Compose (recommended)

```yaml
services:
  stirling-image:
    image: stirlingimage/stirling-image:latest
    container_name: stirling-image
    ports:
      - "1349:1349"
    volumes:
      - stirling-data:/data
      - stirling-workspace:/tmp/workspace
    environment:
      - AUTH_ENABLED=true
      - DEFAULT_USERNAME=admin
      - DEFAULT_PASSWORD=admin
    restart: unless-stopped

volumes:
  stirling-data:
  stirling-workspace:
```

## What's inside the container

Multi-stage Docker build:
1. Build stage — installs Node.js dependencies, builds React frontend with Vite
2. Production stage — Node 22 image with Python 3, ImageMagick, Tesseract, potrace, ML packages, pre-downloaded model weights

Python packages: rembg (BiRefNet-Lite), RealESRGAN, PaddleOCR, MediaPipe, LaMa Cleaner, onnxruntime, opencv-python, Pillow, numpy.

## Health check

```
GET /api/v1/health
```

Returns 200 OK if the server is running. Used by Docker HEALTHCHECK.

## Reverse proxy (nginx example)

```nginx
server {
    listen 80;
    server_name images.example.com;
    client_max_body_size 200M;

    location / {
        proxy_pass http://localhost:1349;
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}
```

---

# REST API

Base URL: `http://localhost:1349`

Interactive API docs on running instance: `/api/docs`
OpenAPI 3.1 spec: `/api/v1/openapi.yaml`

## Authentication

1. Session token — POST /api/auth/login with {username, password}. Returns {token}. Pass as `Authorization: Bearer <token>`.
2. API key — Generate via POST /api/v1/api-keys. Prefixed with `si_`. Pass as `Authorization: Bearer si_...`.

## Tools

### Execute a tool

```
POST /api/v1/tools/:toolId
Content-Type: multipart/form-data
```

Fields: `file` (image), `settings` (JSON string with tool-specific options).

Response: `{ jobId, downloadUrl, originalSize, processedSize }`

### Available tool IDs

Image operations: resize, crop, rotate, convert, compress, strip-metadata, border
Color: brightness-contrast, saturation, color-channels, color-effects, replace-color
Text/codes: watermark-text, watermark-image, text-overlay, qr-generate, barcode-read, ocr
Composition: compose, collage, split, image-to-pdf
Analysis: info, compare, find-duplicates, color-palette
Conversion: svg-to-raster, vectorize, favicon, gif-tools
AI-powered: remove-background, upscale, blur-faces, erase-object, smart-crop
Utility: bulk-rename

### Batch processing

```
POST /api/v1/tools/:toolId/batch
Content-Type: multipart/form-data
```

Multiple files with same settings. Returns ZIP file. X-Job-Id header for progress tracking.

## Pipelines

- Execute: POST /api/v1/pipeline/execute (multipart: file + steps JSON array)
- Save: POST /api/v1/pipeline/save (JSON: {name, description, steps})
- List: GET /api/v1/pipeline/list
- Delete: DELETE /api/v1/pipeline/:id

## Files

- Upload: POST /api/v1/upload
- Download: GET /api/v1/download/:jobId/:filename (public, no auth)
- List saved: GET /api/v1/files
- Save to library: POST /api/v1/files/upload
- Save result: POST /api/v1/files/save-result (JSON: {jobId, filename})
- Get metadata: GET /api/v1/files/:id
- Download saved: GET /api/v1/files/:id/download
- Thumbnail: GET /api/v1/files/:id/thumbnail
- Delete: DELETE /api/v1/files (JSON: {ids: [...]})

## API Keys

- Create: POST /api/v1/api-keys (JSON: {name}) — returns full key once
- List: GET /api/v1/api-keys — metadata only
- Delete: DELETE /api/v1/api-keys/:id

## Settings

- Get all: GET /api/v1/settings
- Get one: GET /api/v1/settings/:key
- Update: PUT /api/v1/settings (admin only, JSON key-value object)

## Teams

- List: GET /api/v1/teams
- Create: POST /api/v1/teams (JSON: {name})
- Rename: PUT /api/v1/teams/:id (JSON: {name})
- Delete: DELETE /api/v1/teams/:id

## Auth endpoints

- Login: POST /api/auth/login
- Logout: POST /api/auth/logout
- Session: GET /api/auth/session
- Change password: POST /api/auth/change-password
- List users (admin): GET /api/auth/users
- Create user (admin): POST /api/auth/register
- Update user (admin): PUT /api/auth/users/:id
- Delete user (admin): DELETE /api/auth/users/:id
- Reset password (admin): POST /api/auth/users/:id/reset-password

## Branding

- Get logo: GET /api/v1/settings/logo (public)
- Upload logo (admin): POST /api/v1/settings/logo
- Delete logo (admin): DELETE /api/v1/settings/logo

## Progress tracking (SSE)

```
GET /api/v1/jobs/:jobId/progress
```

Server-Sent Events stream: {status, progress (0-100), completedFiles, failedFiles}

## Error responses

```json
{ "statusCode": 400, "error": "Bad Request", "message": "Invalid image format" }
```

Status codes: 400 (invalid input), 401 (not authenticated), 403 (not authorized), 413 (file too large), 429 (rate limited), 500 (server error)

---

# Image Engine

The @stirling-image/image-engine package handles all non-AI image operations via Sharp.

## Operations

### resize
width (number), height (number), fit (cover/contain/fill/inside/outside), withoutEnlargement (boolean), percentage (number)

### crop
left (number), top (number), width (number), height (number)

### rotate
angle (number, degrees)

### flip
direction: horizontal or vertical

### convert
format (jpeg/png/webp/avif/tiff/gif), quality (1-100)

### compress
quality (1-100), format (optional override)

### strip-metadata
Removes EXIF, IPTC, XMP metadata. No parameters.

### Color adjustments
brightness (-100 to 100), contrast (-100 to 100), saturation (-100 to 100)

### Color filters
grayscale, sepia, invert — no parameters

### Color channels
red (-100 to 100), green (-100 to 100), blue (-100 to 100)

Supported input formats: JPEG, PNG, WebP, AVIF, TIFF, GIF, SVG, RAW (via libraw)

---

# AI Engine

The @stirling-image/ai package wraps Python ML models. Each operation spawns a Python subprocess. All model weights are bundled in the Docker image.

## Background removal
Model: BiRefNet-Lite via rembg. Parameters: model (string), alphaMatting (boolean), alphaMattingForegroundThreshold (0-255), alphaMattingBackgroundThreshold (0-255).

## Upscaling
Model: RealESRGAN. Parameters: scale (2 or 4).

## OCR
Model: PaddleOCR. Parameters: language (en/ch/fr/de/etc). Returns text with bounding boxes and confidence scores.

## Face detection and blurring
Model: MediaPipe. Parameters: blurStrength (number). Returns blurred image with face region metadata.

## Object erasing (inpainting)
Model: LaMa. Input: image + mask (white = erase, black = keep). Returns inpainted image.

## Bridge mechanism
TypeScript bridge spawns Python subprocess, writes input to temp file, parses JSON progress from stderr for SSE streaming, reads output image, cleans up. Timeout: 5 minutes.