longkeyy/ollama-ui

By longkeyy

•Updated about 1 year ago

Ollama model management platform with API gateway.

Image

Machine learning & AI

2.1K

Overview Tags

longkeyy/ollama-ui repository overview

⁠Ollama Remote Management Platform - Docker Deployment

🚀 Comprehensive Docker deployment guide for Ollama Remote Management Platform

⁠🎯 Overview

This guide provides detailed instructions for deploying Ollama Remote Management Platform using Docker. The platform is designed for remote Ollama model management with secure API access and comprehensive compatibility.

⁠✨ Core Capabilities

⁠🌐 Remote Management

Centralized Control: Manage Ollama models remotely through web interface
Multi-Instance Support: Connect and manage multiple Ollama servers
Real-time Monitoring: Live performance metrics and resource tracking
Secure Access: API key authentication for remote operations

⁠🔌 API Compatibility

OpenAI Compatible: Drop-in replacement for OpenAI API endpoints
Claude Compatible: Optimized for claude-code integration
Unified Interface: Single platform supporting multiple API formats
Request Routing: Intelligent routing based on client requirements

⁠🛡️ Security & Access Control

API Key Management: Secure remote access with customizable keys
Multi-user Support: Role-based permissions and user management
Session Control: Secure authentication and session management
Access Logging: Comprehensive audit trails

⁠🎨 Web Interface

Responsive Design: Optimized for desktop and mobile devices
Real-time Updates: Live streaming chat responses and model status
Multi-language Support: International user interface
Customizable Themes: Dark/light mode with automatic detection

⁠🐳 Docker Image Versions

⁠Standalone Version (Remote Management Only)

Perfect for connecting to existing Ollama servers with minimal resource usage.

# Pull the latest standalone version
docker pull longkeyy/ollama-ui:0.0.1

⁠All-in-One Version (Complete Solution)

Includes both the management platform and Ollama runtime for single-server deployment.

# CUDA/CPU version
docker pull longkeyy/ollama-ui:0.0.1-0.11.8

# AMD ROCM version
docker pull longkeyy/ollama-ui:0.0.1-0.11.8-rocm

⁠🚀 Quick Deployment

⁠Standalone Deployment (Connect to Existing Ollama)

Perfect for users who already have Ollama running elsewhere:

# Basic standalone deployment
docker run -d \
  --name ollama-remote \
  -p 8083:8083 \
  -v ollama-data:/app/data \
  longkeyy/ollama-ui:0.0.1

# Connect to remote Ollama server
docker run -d \
  --name ollama-remote \
  -p 8083:8083 \
  -e OLLAMA_API_BASE_URL=http://your-ollama-server:11434 \
  -e API_KEY_REQUIRED=true \
  -v ollama-data:/app/data \
  longkeyy/ollama-ui:0.0.1

⁠All-in-One Deployment (Complete Solution)

Recommended for new users who want everything in one container. The All-in-One version exposes the web interface on port 11434, making it a drop-in replacement for standard Ollama:

# CPU version
docker run -d \
  --name ollama-complete \
  -p 11434:11434 \
  -v ollama-models:/root/.ollama \
  -v ollama-data:/app/data \
  longkeyy/ollama-ui:0.0.1-0.11.8

# GPU version (NVIDIA CUDA)
docker run -d \
  --name ollama-complete-gpu \
  --gpus all \
  -p 11434:11434 \
  -v ollama-models:/root/.ollama \
  -v ollama-data:/app/data \
  longkeyy/ollama-ui:0.0.1-0.11.8

# AMD ROCM version
docker run -d \
  --name ollama-complete-rocm \
  --device=/dev/kfd --device=/dev/dri \
  --security-opt seccomp=unconfined \
  -p 11434:11434 \
  -v ollama-models:/root/.ollama \
  -v ollama-data:/app/data \
  longkeyy/ollama-ui:0.0.1-0.11.8-rocm

⁠📋 Docker Compose Examples

⁠Standalone with External Ollama

version: '3.8'

services:
  ollama:
    image: ollama/ollama:latest
    container_name: ollama-server
    ports:
      - "11434:11434"
    volumes:
      - ollama-models:/root/.ollama
    environment:
      - OLLAMA_HOST=0.0.0.0:11434
    restart: unless-stopped

  ollama-ui:
    image: longkeyy/ollama-ui:0.0.1
    container_name: ollama-remote-ui
    ports:
      - "8083:8083"
    volumes:
      - ollama-data:/app/data
    environment:
      - OLLAMA_API_BASE_URL=http://ollama:11434
      - API_KEY_REQUIRED=true
      - DEFAULT_MODEL=llama2:7b
    depends_on:
      - ollama
    restart: unless-stopped

volumes:
  ollama-models:
  ollama-data:

⁠All-in-One Complete Solution

version: '3.8'

services:
  ollama-complete:
    image: longkeyy/ollama-ui:0.0.1-0.11.8
    container_name: ollama-complete
    ports:
      - "11434:11434"  # Single port for both API and Web UI
    volumes:
      - ollama-models:/root/.ollama
      - ollama-data:/app/data
    environment:
      - OLLAMA_HOST=0.0.0.0:11434
      - OLLAMA_UI_PORT=11434  # WebUI also uses port 11434
      - API_KEY_REQUIRED=true
    restart: unless-stopped
    # GPU support (uncomment for NVIDIA GPU)
    # deploy:
    #   resources:
    #     reservations:
    #       devices:
    #         - driver: nvidia
    #           count: all
    #           capabilities: [gpu]

volumes:
  ollama-models:
  ollama-data:

⁠⚙️ Environment Variables

⁠Core Configuration

All configuration options can be set via environment variables with the WEBUI_ prefix:

Variable	Default	Description
`WEBUI_SERVER_PORT`	`8080` (standalone) / `11434` (AIO)	Web interface port
`WEBUI_SERVER_HOST`	`0.0.0.0`	Server listening address
`WEBUI_OLLAMA_URL`	`http://localhost:11434`	Ollama API endpoint
`WEBUI_OLLAMA_TIMEOUT`	`300s`	Ollama request timeout

⁠Authentication Settings

Variable	Default	Description
`WEBUI_AUTH_USERNAME`	`admin`	Default login username
`WEBUI_AUTH_PASSWORD`	`admin123`	Default login password (⚠️ Change immediately!)
`WEBUI_AUTH_JWT_SECRET`	`your-secret-key-change-in-production`	JWT secret key
`WEBUI_AUTH_TOKEN_EXPIRE`	`3600`	Token expiration time (seconds)

⁠Download Manager Settings

Variable	Default	Description
`WEBUI_DOWNLOADS_MAX_CONCURRENT`	`3`	Max concurrent downloads
`WEBUI_DOWNLOADS_QUEUE_SIZE`	`10`	Download queue size
`WEBUI_DOWNLOADS_RETRY_ATTEMPTS`	`3`	Retry attempts on failure
`WEBUI_DOWNLOADS_RETRY_DELAY`	`5s`	Delay between retries

⁠Logging Configuration

Variable	Default	Description
`WEBUI_LOGGING_LEVEL`	`info`	Log level (debug, info, warn, error)
`WEBUI_LOGGING_FORMAT`	`json`	Log format (json, text)
`WEBUI_LOGGING_FILE`	`ollama-webui.log`	Log file path

⁠Ollama Configuration (All-in-One Version)

In the All-in-One version, Ollama runs internally and is not exposed:

Variable	Default	Description
`OLLAMA_HOST`	`127.0.0.1:11435`	Ollama internal address (protected)
`OLLAMA_MODELS`	`/root/.ollama/models`	Model storage directory
`OLLAMA_PRELOAD_MODELS`	(optional)	Comma-separated models to preload

⁠🔧 Data Persistence

⁠Important Directories

/root/.ollama: Ollama models and configuration (All-in-One version)
/app/data: Platform data, user settings, API keys, chat history
/var/log/supervisor: Service logs (All-in-One version)

⁠Backup and Restore

# Backup platform data
docker run --rm -v ollama-data:/source -v $(pwd):/backup alpine \
  tar czf /backup/platform-data-backup.tar.gz -C /source .

# Backup models (All-in-One version)
docker run --rm -v ollama-models:/source -v $(pwd):/backup alpine \
  tar czf /backup/models-backup.tar.gz -C /source .

# Restore platform data
docker run --rm -v ollama-data:/target -v $(pwd):/backup alpine \
  tar xzf /backup/platform-data-backup.tar.gz -C /target

# Restore models
docker run --rm -v ollama-models:/target -v $(pwd):/backup alpine \
  tar xzf /backup/models-backup.tar.gz -C /target

⁠🌐 Access and Usage

⁠Initial Setup

Access the Platform:
- Standalone: http://localhost:8083
- All-in-One: http://localhost:11434
Default Login Credentials:
- Username: admin
- Password: admin123
⚠️ IMPORTANT: Please change the default password immediately after first login!

Change Default Password:

# Set new credentials via environment variables
docker run -d \
  -e AUTH_USERNAME=your_username \
  -e AUTH_PASSWORD=your_secure_password \
  # ...other parameters

Or via configuration file config.yaml:

auth:
  username: "your_username"
  password: "your_secure_password"

API Key Setup: Configure API keys for remote access
Model Management: Install your first models

⁠API Endpoints

OpenAI Compatible:

GET /v1/models - List available models
POST /v1/chat/completions - Chat completions
POST /v1/embeddings - Text embeddings

Claude Compatible:

POST /v1/claude/messages - Claude-style messaging
GET /v1/claude/models - Claude model list

⁠Usage Examples

# OpenAI-style chat
curl -X POST http://localhost:8083/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
    "model": "llama2:7b",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": false
  }'

# Claude-style messaging
curl -X POST http://localhost:8083/v1/claude/messages \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
    "model": "llama2:7b",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

⁠🔍 Troubleshooting

⁠Common Issues

1. Cannot access the web interface

# Check container status
docker ps -a
docker logs ollama-remote

# Verify port accessibility
netstat -tulpn | grep :8083

2. API authentication failures

# Check API key configuration
docker exec ollama-remote cat /app/data/config.yaml

# Verify API key in requests
curl -H "Authorization: Bearer your-key" http://localhost:8083/v1/models

3. Ollama connection issues

# Test Ollama connectivity (All-in-One)
curl http://localhost:11434/api/version

# Test external Ollama (Standalone)
docker exec ollama-remote curl $OLLAMA_API_BASE_URL/api/version

4. Model download problems

# Check available disk space
docker exec ollama-remote df -h

# Monitor download progress
docker logs -f ollama-remote

⁠Performance Optimization

Memory Management:

# Limit container memory
docker run -d --memory=8g --memory-swap=12g \
  # ... other parameters

CPU Allocation:

# Limit CPU usage
docker run -d --cpus=4.0 --cpu-shares=2048 \
  # ... other parameters

GPU Optimization:

# Check GPU availability
docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi

# Monitor GPU usage
docker exec ollama-complete nvidia-smi

⁠📊 Monitoring and Logging

⁠Health Checks

# Platform health check
curl -f http://localhost:8083/health || echo "Platform unhealthy"

# API health check
curl -f http://localhost:8083/v1/models || echo "API unhealthy"

# Ollama service health (All-in-One)
curl -f http://localhost:11434/api/version || echo "Ollama unhealthy"

⁠Log Management

# View platform logs
docker logs -f ollama-remote

# View specific service logs (All-in-One)
docker exec ollama-complete supervisorctl status
docker exec ollama-complete tail -f /var/log/supervisor/ollama.log

⁠🔄 Updates and Maintenance

⁠Update to Latest Version

# Stop current container
docker stop ollama-remote
docker rm ollama-remote

# Pull latest image
docker pull longkeyy/ollama-ui:0.0.1-0.11.8

# Restart with same configuration
docker run -d \
  --name ollama-remote \
  # ... same parameters as before
  longkeyy/ollama-ui:0.0.1-0.11.8

⁠Version Migration

# Export current configuration
docker exec ollama-remote cat /app/data/config.yaml > config-backup.yaml

# Import to new version
docker cp config-backup.yaml ollama-remote:/app/data/config.yaml

⁠🌟 Advanced Features

⁠API Gateway Configuration

# Enable load balancing across multiple Ollama instances
docker run -d \
  --name ollama-gateway \
  -p 8083:8083 \
  -e OLLAMA_BACKENDS="http://ollama1:11434,http://ollama2:11434" \
  -e LOAD_BALANCE_STRATEGY="round_robin" \
  longkeyy/ollama-ui:0.0.1

⁠Multi-Architecture Support

# Explicitly pull for specific architecture
docker pull --platform linux/amd64 longkeyy/ollama-ui:0.0.1-0.11.8
docker pull --platform linux/arm64 longkeyy/ollama-ui:0.0.1-0.11.8

⁠🏗️ Architecture

⁠All-in-One Architecture (Protected Ollama)

┌───────────────────────────────────────────┐
│        ollama-allinone Container            │
│                                              │
│  ┌─────────────────┐  ┌────────────────┐ │
│  │ Ollama (Internal) │←─│ WebUI (Public)  │ │
│  │ 127.0.0.1:11435   │  │ 0.0.0.0:11434   │ │
│  └─────────────────┘  └────────────────┘ │
│         ↑                        ↓           │
│         └──── Protected ───────┘           │
│                                              │
│      Supervisor Process Manager              │
└───────────────────────────────────────────┘
        External Access: Port 11434 Only

⁠Standalone Architecture

┌──────────────┐     ┌──────────────┐
│ ollama       │     │ ollama-ui    │
│ Container    │←────│ Container    │
│ :11434       │     │ :8083        │
└──────────────┘     └──────────────┘

⁠🤝 Support & Community

GitHub Repository: longkeyy/ollama-ui⁠
Docker Hub: longkeyy/ollama-ui⁠
Issues & Support: GitHub Issues⁠

💡 Ready to deploy? Start with our recommended All-in-One deployment for the complete experience!

Tag summary

Recent tags

Content type

Image

Digest

sha256:72e7f4bcc…

Size

1.1 GB

Last updated

about 1 year ago

docker pull longkeyy/ollama-ui