Use Open WebUI with Eden AI - Eden AI Documentation

Configure Open WebUI (formerly Ollama WebUI) to use Eden AI for accessing 500+ AI models through a sleek interface.

Overview

Open WebUI is a self-hosted, feature-rich web interface for AI models. By connecting it to Eden AI, you get:

500+ models: Access OpenAI, Anthropic, Google, Cohere, and more
Modern UI: Beautiful, responsive interface similar to ChatGPT
RAG support: Document upload and retrieval-augmented generation
Multi-user: User management with authentication
Self-hosted: Complete data privacy and control

Prerequisites

Docker installed
Eden AI API key from https://app.edenai.run
2GB+ RAM recommended

Quick Start

Option 1: Docker (Recommended)

Run Open WebUI with a single command:

docker run -d \
  --name open-webui \
  -p 3000:8080 \
  -v open-webui:/app/backend/data \
  -e OPENAI_API_BASE_URL=https://api.edenai.run/v3 \
  -e OPENAI_API_KEY=YOUR_EDEN_AI_API_KEY \
  -e WEBUI_AUTH=False `# Disable authentication for single-user setup` \
  ghcr.io/open-webui/open-webui:main

Access Open WebUI at http://localhost:3000

Option 2: Docker Compose

Create a docker-compose.yml file:

version: '3.8'

services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    ports:
      - "3000:8080"
    environment:
      - OPENAI_API_BASE_URL=https://api.edenai.run/v3
      - OPENAI_API_KEY=${EDEN_AI_API_KEY}
      - WEBUI_SECRET_KEY=${WEBUI_SECRET_KEY}
    volumes:
      - open-webui:/app/backend/data
    restart: unless-stopped

volumes:
  open-webui:

EDEN_AI_API_KEY=your_eden_ai_api_key_here
WEBUI_SECRET_KEY=your_random_secret_key_here

# Start Open WebUI
docker compose up -d

# View logs
docker compose logs -f

# Stop
docker compose down

Configuration

Initial Setup

Open browser: Navigate to http://localhost:3000
Create admin account: Register the first user (becomes admin)
Configure models: Go to Settings → Models

Add Eden AI Models

In the Open WebUI interface:

Click Settings (gear icon)
Go to Models tab
Add models in the format provider/model:

Environment Variables

Configure Open WebUI with environment variables:

# Eden AI Configuration
OPENAI_API_BASE_URL=https://api.edenai.run/v3
OPENAI_API_KEY=your_eden_ai_api_key

# Application Settings
WEBUI_SECRET_KEY=your_secret_key_here
WEBUI_NAME=Eden AI Chat
DATA_DIR=/app/backend/data

# Authentication
ENABLE_SIGNUP=true
DEFAULT_USER_ROLE=user

# Optional: RAG Settings
ENABLE_RAG_WEB_SEARCH=true
RAG_EMBEDDING_MODEL=text-embedding-3-small
RAG_TOP_K=5
CHUNK_SIZE=1500
CHUNK_OVERLAP=100

# Optional: Image Generation
ENABLE_IMAGE_GENERATION=true
IMAGE_GENERATION_MODEL=openai/gpt-image-2

Features

Chat with Multiple Models

Select model: Click the model dropdown at the top
Start chatting: Type your message
Switch models: Change models mid-conversation
Compare responses: Use split-screen to compare model outputs

Document Upload (RAG)

Upload documents for context-aware conversations:

services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    environment:
      - OPENAI_API_BASE_URL=https://api.edenai.run/v3
      - OPENAI_API_KEY=${EDEN_AI_API_KEY}
      # Enable RAG
      - ENABLE_RAG_WEB_SEARCH=true
      - RAG_EMBEDDING_MODEL=openai/text-embedding-3-small
      - RAG_EMBEDDING_ENGINE=openai
      - RAG_EMBEDDING_OPENAI_BATCH_SIZE=1

Usage:

Click the + icon in chat
Upload PDF, DOCX, TXT, or other documents
Ask questions about the uploaded content
The AI retrieves relevant sections automatically

Image Generation

Generate images through Eden AI’s image models:

# Enable image generation
ENABLE_IMAGE_GENERATION=true
IMAGE_GENERATION_MODEL=openai/gpt-image-2
IMAGE_GENERATION_ENGINE=openai

# Use Eden AI endpoint
AUTOMATIC1111_BASE_URL=https://api.edenai.run/v3

Usage: Type /imagine followed by your prompt in the chat.

Web Search

Open WebUI’s Perplexity Search engine works with Eden AI’s Perplexity-compatible search API. The engine sends a fixed body with no model, so the provider goes in the URL path:

Open Admin Panel, then Settings, then Web Search.
Turn on Enable Web Search and select perplexity_search as the Web Search Engine.
Enter your Eden AI API key as the API key.
Set Perplexity Search API URL to https://api.edenai.run/v3/search/tavily (or end it with another provider id from GET /v3/search/models).

On a new install, you can set the same values through environment variables instead:

ENABLE_WEB_SEARCH=true
WEB_SEARCH_ENGINE=perplexity_search
PERPLEXITY_API_KEY=your_eden_ai_api_key
PERPLEXITY_SEARCH_API_URL=https://api.edenai.run/v3/search/tavily

Open WebUI reads these variables only on its first start and saves them in its database. From then on the saved values win, so on an existing install change them in the Admin Panel (or set ENABLE_PERSISTENT_CONFIG=false to make Open WebUI read the environment on every start).

Keep the search result count (WEB_SEARCH_RESULT_COUNT, default 3) at 20 or below. Eden AI rejects a larger max_results, and Open WebUI hides the error: searches return no results and no message says why.

Each query Open WebUI sends is a billed search on your Eden AI account, and Open WebUI can generate more than one query for a single chat message. GET /v3/search/models lists each provider’s price.

Voice Input

Enable voice-to-text:

ENABLE_AUDIO=true
AUDIO_STT_ENGINE=openai
AUDIO_STT_MODEL=whisper-1

Advanced Configuration

Multiple API Endpoints

Configure multiple providers:

services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    environment:
      # Eden AI as primary
      - OPENAI_API_BASE_URL=https://api.edenai.run/v3
      - OPENAI_API_KEY=${EDEN_AI_API_KEY}

      # Additional endpoints (optional)
      - OPENAI_API_BASE_URLS=https://api.edenai.run/v3,https://api.openai.com/v1
      - OPENAI_API_KEYS=${EDEN_AI_API_KEY},${OPENAI_API_KEY}

Custom Model Metadata

Define model capabilities and pricing:

{
  "models": [
    {
      "id": "anthropic/claude-sonnet-latest",
      "name": "Claude Sonnet",
      "owned_by": "anthropic",
      "created": 1729900800,
      "object": "model",
      "info": {
        "description": "Most intelligent Claude model",
        "capabilities": {
          "vision": true,
          "function_calling": true
        }
      }
    },
    {
      "id": "openai/gpt-latest",
      "name": "GPT",
      "owned_by": "openai",
      "created": 1715300000,
      "object": "model",
      "info": {
        "description": "GPT-4 with vision",
        "capabilities": {
          "vision": true,
          "function_calling": true
        }
      }
    }
  ]
}

User Permissions

Configure role-based access:

# User registration
ENABLE_SIGNUP=true
DEFAULT_USER_ROLE=user

# Admin settings
ADMIN_EMAIL=admin@example.com

# Model access control
ENABLE_MODEL_FILTER=true
MODEL_FILTER_LIST=anthropic/claude-sonnet-latest,openai/gpt-latest

Persistent Storage

Configure data persistence:

services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    volumes:
      # Data persistence
      - ./data:/app/backend/data
      # Custom models
      - ./models.json:/app/backend/models.json:ro
      # User uploads
      - ./uploads:/app/backend/uploads

Security

Authentication

Enable and configure authentication:

# Require login
WEBUI_AUTH=true

# Session settings
WEBUI_SECRET_KEY=generate_strong_random_key_here
SESSION_TIMEOUT=3600

# OAuth (optional)
ENABLE_OAUTH_SIGNUP=true
OAUTH_CLIENT_ID=your_client_id
OAUTH_CLIENT_SECRET=your_client_secret
OAUTH_PROVIDER_NAME=Google

HTTPS Setup

Use a reverse proxy for HTTPS:

server {
  listen 443 ssl http2;
  server_name chat.yourdomain.com;

  ssl_certificate /etc/ssl/certs/cert.pem;
  ssl_certificate_key /etc/ssl/private/key.pem;

  location / {
    proxy_pass http://localhost:3000;
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Proto $scheme;

    # WebSocket support
    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";
  }
}

Rate Limiting

Protect your API key:

# Rate limiting
ENABLE_RATE_LIMIT=true
RATE_LIMIT_REQUESTS=60
RATE_LIMIT_WINDOW=60

# Per-user limits
USER_PERMISSIONS_CHAT_DAILY_LIMIT=100

Production Deployment

Full Production Stack

version: '3.8'

services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    restart: always
    ports:
      - "127.0.0.1:3000:8080"
    environment:
      - OPENAI_API_BASE_URL=https://api.edenai.run/v3
      - OPENAI_API_KEY=${EDEN_AI_API_KEY}
      - WEBUI_SECRET_KEY=${WEBUI_SECRET_KEY}
      - ENABLE_SIGNUP=false
      - DEFAULT_USER_ROLE=user
      - WEBUI_AUTH=true
    volumes:
      - ./data:/app/backend/data
      - ./uploads:/app/backend/uploads
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8080/health"]
      interval: 30s
      timeout: 10s
      retries: 3

  nginx:
    image: nginx:alpine
    container_name: nginx-proxy
    restart: always
    ports:
      - "80:80"
      - "443:443"
    volumes:
      - ./nginx.conf:/etc/nginx/nginx.conf:ro
      - ./ssl:/etc/ssl:ro
    depends_on:
      - open-webui

Backup Strategy

Backup your data regularly:

#!/bin/bash

# Backup directory
BACKUP_DIR="./backups/$(date +%Y%m%d_%H%M%S)"
mkdir -p "$BACKUP_DIR"

# Backup data
docker compose exec open-webui tar czf /tmp/backup.tar.gz /app/backend/data
docker cp open-webui:/tmp/backup.tar.gz "$BACKUP_DIR/"

# Backup database (if using external DB)
# docker compose exec postgres pg_dump -U user dbname > "$BACKUP_DIR/db.sql"

echo "Backup completed: $BACKUP_DIR"

Troubleshooting

Models Not Appearing

If models don’t show up:

Check API key:

docker compose logs open-webui | grep -i "api"

Verify base URL:

docker compose exec open-webui env | grep OPENAI_API_BASE_URL

Test endpoint manually:

curl -X POST https://api.edenai.run/v3/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-latest",
    "messages": [{"role": "user", "content": "test"}],
    "stream": true
  }'

Restart container:
docker compose restart open-webui

Authentication Issues

If you can’t log in:

Reset admin password:

docker compose exec open-webui python manage.py reset-admin-password

Check secret key:

docker compose exec open-webui env | grep WEBUI_SECRET_KEY

Performance Issues

If the UI is slow:

Check resource usage:
docker stats open-webui

Increase memory:

services:
  open-webui:
    deploy:
      resources:
        limits:
          memory: 4G

Optimize database:

docker compose exec open-webui python manage.py optimize-db

RAG Not Working

If document upload fails:

Check volume mounts:

docker compose exec open-webui ls -la /app/backend/uploads

Verify embedding model:

docker compose exec open-webui env | grep RAG_EMBEDDING

Check logs:

docker compose logs open-webui | grep -i "rag"

Cost Optimization

1. Use Appropriate Models

Configure cheaper models for embeddings:

# Use smaller embedding model
RAG_EMBEDDING_MODEL=text-embedding-3-small

# Use GPT-3.5 for simple tasks
DEFAULT_MODEL=openai/gpt-mini-latest

2. Limit Token Usage

Set maximum tokens:

# Limit response length
MAX_TOKENS=1000

# Limit context
MAX_CONTEXT_LENGTH=4000

3. Monitor Usage

Track costs in Eden AI dashboard:

Visit https://app.edenai.run
View usage reports
Set billing alerts

Example Workflows

1. Customer Support Bot

Upload support docs using RAG
Create templates for common queries
Use Claude Haiku for cost-effective responses

2. Code Assistant

Use GPT-4 or Claude for complex code
Enable file upload for code review
Configure presets for different languages

3. Research Assistant

Enable web search for current information
Use Gemini Pro for long context
RAG for internal documents

Next Steps

LibreChat - Alternative chat platform
Continue.dev - IDE integration
Monitoring - Track usage and spending

​Overview

​Prerequisites

​Quick Start

​Option 1: Docker (Recommended)

​Option 2: Docker Compose

​Configuration

​Initial Setup

​Add Eden AI Models

​Environment Variables

​Features

​Chat with Multiple Models

​Document Upload (RAG)

​Image Generation

​Web Search

​Voice Input

​Advanced Configuration

​Multiple API Endpoints

​Custom Model Metadata

​User Permissions

​Persistent Storage

​Security

​Authentication

​HTTPS Setup

​Rate Limiting

​Production Deployment

​Full Production Stack

​Backup Strategy

​Troubleshooting

​Models Not Appearing

​Authentication Issues

​Performance Issues

​RAG Not Working

​Cost Optimization

​1. Use Appropriate Models

​2. Limit Token Usage

​3. Monitor Usage

​Example Workflows

​1. Customer Support Bot

​2. Code Assistant

​3. Research Assistant

​Next Steps

Overview

Prerequisites

Quick Start

Option 1: Docker (Recommended)

Option 2: Docker Compose

Configuration

Initial Setup

Add Eden AI Models

Environment Variables

Features

Chat with Multiple Models

Document Upload (RAG)

Image Generation

Web Search

Voice Input

Advanced Configuration

Multiple API Endpoints

Custom Model Metadata

User Permissions

Persistent Storage

Security

Authentication

HTTPS Setup

Rate Limiting

Production Deployment

Full Production Stack

Backup Strategy

Troubleshooting

Models Not Appearing

Authentication Issues

Performance Issues

RAG Not Working

Cost Optimization

1. Use Appropriate Models

2. Limit Token Usage

3. Monitor Usage

Example Workflows

1. Customer Support Bot

2. Code Assistant

3. Research Assistant

Next Steps