Shrike Media Server

⚠️ Alpha Software Warning: Shrike Media Server is in the alpha phase of development. The server binds to port 8090 and exposes your media library over HTTP with no authentication. If you don't want your media accessible on your network, ensure the port is blocked by your firewall or only run Shrike on trusted networks.

Shrike Media Server is a companion for the Lowkey Media Viewer. It allows for managing long-running offline tasks like media tagging, transcript generation, media conversion, file serving, and media ingestion.

Features

Job Queue Management: Create, monitor, cancel, and manage long-running media processing jobs
Media Browser: Browse, search, and preview media files with a modern web interface
Auto-Tagging: Automatic image tagging using ONNX-based machine learning models (WD Tagger)
LLM Descriptions: Generate image descriptions using Ollama vision models
Transcription: Generate video transcripts using Faster Whisper
Media Processing: FFmpeg integration for media conversion and manipulation
Media Ingestion: Batch import media files into the database
Real-Time Updates: Server-Sent Events (SSE) for live job status updates
System Tray: Windows system tray integration for easy access
Database Persistence: SQLite-backed job and media database

System Requirements

Operating System: Windows 10/11 (x64)
Disk Space: ~500MB for binaries and embedded tools
RAM: 4GB minimum, 8GB+ recommended for ML tagging

Installation (End Users)

Run the shrike.exe binary. This will start the background server and create an icon in your Windows system tray for launching the WebUI.
Check Lowkey Database Path. Open the Config tab in the Web UI. You should see the path to your Lowkey Media Database. If it's incorrect, you can manually change it here.
Set Up Model Paths. You'll need to point the server at a few files to enable auto tagging:

Development Setup

Prerequisites

1. Install Go

Shrike requires Go 1.24.0 or later.

Windows (Recommended):

Download the latest Go installer from https://go.dev/dl/
Run the MSI installer

Verify installation:

go version
# Should output: go version go1.24.x windows/amd64

Alternative (using winget):

winget install GoLang.Go

Alternative (using Chocolatey):

choco install golang

2. Install Git

Required for cloning the repository and Go module management.

winget install Git.Git

3. C Compiler (for CGO - SQLite)

The SQLite package (modernc.org/sqlite) is a pure Go implementation and does not require CGO. No C compiler is needed.

4. Optional: External Tools

For full functionality, install these optional tools:

Tool	Purpose	Installation
Ollama	LLM-based image descriptions	Download and run installer
ONNX Runtime	ML model inference for auto-tagging	Download DLL, configure path in settings
Faster Whisper	Video transcription	Download, configure path in settings

Note: FFmpeg, yt-dlp, gallery-dl, and ExifTool are embedded in the binary and extracted at runtime.

Building from Source

1. Clone the Repository

git clone https://github.com/stevecastle/shrike.git
cd shrike

2. Download Dependencies

go mod download

3. Verify Dependencies

go mod verify

4. Build the Executable

# Standard build
go build -o shrike.exe .

# Build with optimizations (smaller binary, no debug info)
go build -ldflags="-s -w" -o shrike.exe .

5. Run the Server

.\shrike.exe

The server will start on http://localhost:8090 and a system tray icon will appear.

Development Mode

For faster iteration during development, you can use go run:

go run .

Running Tests

# Run all tests
go test ./...

# Run tests with verbose output
go test -v ./...

# Run tests for a specific package
go test -v ./media/...

Project Structure

shrike/
├── main.go                 # Application entry point, HTTP handlers, system tray
├── go.mod                  # Go module definition
├── go.sum                  # Dependency checksums
│
├── appconfig/              # Application configuration management
│   └── config.go           # Config loading/saving, defaults
│
├── assets/                 # Static assets
│   └── logo.ico            # System tray icon
│
├── client/                 # Frontend static files
│   └── static/
│       ├── styles.css      # Main stylesheet
│       └── details.css     # Job details page styles
│
├── embedexec/              # Embedded executable extraction
│   ├── embedexec.go        # Runtime extraction of embedded binaries
│   └── bin/                # Embedded executables (see below)
│
├── jobqueue/               # Job queue implementation
│   └── jobqueue.go         # Queue, job state, persistence
│
├── media/                  # Media database operations
│   ├── media.go            # Media queries, search, pagination
│   └── media_test.go       # Media tests
│
├── onnxtag/                # ONNX-based image tagging
│   ├── onnxtag.go          # Tagger implementation
│   └── config.go           # Tagger configuration
│
├── renderer/               # HTML template rendering
│   ├── renderer.go         # Template loading, middleware
│   └── templates/          # Go HTML templates
│       ├── index.go.html   # Home page
│       ├── jobs.go.html    # Jobs list
│       ├── detail.go.html  # Job detail view
│       ├── media.go.html   # Media browser
│       ├── config.go.html  # Configuration page
│       └── ...
│
├── runners/                # Job execution runners
│   └── runners.go          # Worker pool implementation
│
├── stream/                 # Server-Sent Events (SSE)
│   └── stream.go           # Real-time update streaming
│
├── tasks/                  # Task implementations
│   ├── registry.go         # Task registration
│   ├── autotag.go          # ONNX auto-tagging
│   ├── autotag_vision.go   # LLM vision tagging
│   ├── command.go          # Generic command execution
│   ├── ffmpeg.go           # FFmpeg processing
│   ├── media_ingest.go     # Media file ingestion
│   ├── media_metadata.go   # Metadata generation
│   ├── media_move.go       # File moving with DB update
│   ├── media_cleanup.go    # Orphan cleanup
│   └── ...
│
├── chrome-extension/       # Browser extension for sending URLs
│   ├── manifest.json
│   ├── popup.html
│   ├── popup.js
│   └── popup.css
│
├── cmd/                    # Additional command-line tools
│   ├── onnxtag/            # Standalone ONNX tagger CLI
│   ├── dbcopy/             # Database copy utility
│   └── sbs/                # Side-by-side comparison tool
│
├── API_DOCUMENTATION.md    # Detailed API reference
└── README.md               # This file

Embedded Binaries

Shrike embeds several third-party executables that are extracted to %ProgramData%\Shrike\tmp\ at runtime:

Binary	Purpose
`ffmpeg.exe`	Media conversion and processing
`ffprobe.exe`	Media file analysis
`ffplay.exe`	Media playback
`yt-dlp.exe`	Video downloading
`gallery-dl.exe`	Image gallery downloading
`exiftool.exe`	Metadata extraction
`faster-whisper-xxl.exe`	Video transcription (optional)
`dedupe.exe`	Duplicate detection
`createdump.exe`	Crash dump creation

The embedded binaries are located in embedexec/bin/. To update them:

Download the new binary
Replace the file in embedexec/bin/
Rebuild the project

Note: Including these binaries significantly increases the final executable size (~200MB+).

Configuration

Configuration is stored at: %APPDATA%\Lowkey Media Viewer\config.json

Configuration Options

{
  "dbPath": "C:\\path\\to\\your\\database.db",

  "ollamaBaseUrl": "http://localhost:11434",
  "ollamaModel": "llama3.2-vision",
  "describePrompt": "Please describe this image...",
  "autotagPrompt": "Please analyze this image and select tags...",

  "onnxTagger": {
    "modelPath": "C:\\path\\to\\model.onnx",
    "labelsPath": "C:\\path\\to\\selected_tags.csv",
    "configPath": "C:\\path\\to\\config.json",
    "ortSharedLibraryPath": "C:\\path\\to\\onnxruntime.dll",
    "generalThreshold": 0.35,
    "characterThreshold": 0.85
  },

  "fasterWhisperPath": "C:\\path\\to\\faster-whisper-xxl.exe"
}

ONNX Tagger Setup

Download model files from HuggingFace:
- model.onnx - The neural network model
- selected_tags.csv - Tag labels
- config.json - Model configuration (optional)
Download ONNX Runtime:
- Get onnxruntime-win-x64-*.zip
- Extract onnxruntime.dll from lib/ folder
Configure paths in the Shrike Config tab

Usage

Once the server is installed, Lowkey Media Viewer should detect it and be able to create jobs. You can also create bulk jobs on the results of any search from the Media Browser tab.

Web Interface

Access the web interface at: http://localhost:8090

Home: Quick job creation
Jobs: View and manage all jobs
Media: Browse and search media files
Config: Configure server settings
Stats: View database statistics

System Tray

Right-click the Shrike icon in the system tray:

Open Web UI: Launch browser to web interface
Quit: Shutdown the server

Available Tasks

Task ID	Name	Description
`wait`	Wait	Test task that waits 5 seconds
`gallery-dl`	gallery-dl	Download media from websites
`yt-dlp`	yt-dlp	Download videos from YouTube, etc.
`ffmpeg`	ffmpeg	Process media files
`ingest`	Ingest Media Files	Scan directories and add to database
`metadata`	Generate Metadata	Generate descriptions, hashes, dimensions
`autotag`	Auto Tag (ONNX)	Automatic image tagging with ML
`move`	Move Media Files	Move files and update database
`remove`	Remove Media	Remove entries from database
`cleanup`	CleanUp	Remove orphaned database entries

API Documentation

See API_DOCUMENTATION.md for detailed HTTP API reference including:

Endpoint specifications
Request/response formats
SSE streaming events
Example curl commands

Chrome Extension

A Chrome extension is included for quickly sending URLs to Shrike.

Installation

Open Chrome and go to chrome://extensions/
Enable "Developer mode"
Click "Load unpacked"
Select the chrome-extension/ folder

Usage

Navigate to a page you want to download
Click the Shrike extension icon
Select the download method (yt-dlp, gallery-dl, etc.)
Job is created automatically

Troubleshooting

Port 8090 Already in Use

If another application is using port 8090, you'll need to modify the port in main.go:

srv = &http.Server{
    Addr:    ":8090",  // Change this port
    Handler: mux,
}

Database Connection Errors

Ensure the database path in config.json exists
Check file permissions on the database file
Ensure no other process has locked the database

ONNX Tagger Not Working

Verify all model files are downloaded
Check ONNX Runtime DLL path is correct
Ensure model is compatible (ONNX opset 11+)
Check Windows Event Viewer for crash logs

System Tray Icon Not Appearing

Check Windows notification area settings
Ensure hidden icons are visible
Try restarting Windows Explorer

Contributing

Fork the repository
Create a feature branch
Make your changes
Run tests: go test ./...
Submit a pull request

License

See LICENSE for details.

Name		Name	Last commit message	Last commit date
Latest commit History 91 Commits
.claude		.claude
appconfig		appconfig
assets		assets
chrome-extension		chrome-extension
client/static		client/static
cmd		cmd
deps		deps
embedexec		embedexec
firefox-extension		firefox-extension
jobqueue		jobqueue
media		media
onnxtag		onnxtag
renderer		renderer
runners		runners
stream		stream
tasks		tasks
.gitignore		.gitignore
API_DOCUMENTATION.md		API_DOCUMENTATION.md
CLAUDE.md		CLAUDE.md
LICENSE		LICENSE
README.md		README.md
STREAMING_IMPROVEMENTS.md		STREAMING_IMPROVEMENTS.md
go.mod		go.mod
go.sum		go.sum
main.go		main.go

Folders and files

Latest commit

History

Repository files navigation

Shrike Media Server

Table of Contents

Features

System Requirements

Installation (End Users)

Development Setup

Prerequisites

1. Install Go

2. Install Git

3. C Compiler (for CGO - SQLite)

4. Optional: External Tools

Building from Source

1. Clone the Repository

2. Download Dependencies

3. Verify Dependencies

4. Build the Executable

5. Run the Server

Development Mode

Running Tests

Project Structure

Embedded Binaries

Configuration

Configuration Options

ONNX Tagger Setup

Usage

Web Interface

System Tray

Available Tasks

API Documentation

Chrome Extension

Installation

Usage

Troubleshooting

Port 8090 Already in Use

Database Connection Errors

ONNX Tagger Not Working

System Tray Icon Not Appearing

Contributing

License

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages