Skip to content
DeepDiagramDeepDiagram
Back to Docs

Getting Started

Get DeepDiagram running in 5 minutes

Prerequisites

  • Docker and Docker Compose installed
  • An LLM API key (OpenAI, DeepSeek, or any OpenAI-compatible provider)

Step 1: Clone the Repository

git clone https://github.com/twwch/DeepDiagram.git
cd DeepDiagram

Step 2: Configure Environment Variables

Create a .env file in the project root:

# Choose an LLM provider

# --- Option 1: OpenAI ---
OPENAI_API_KEY=sk-proj-xxxxxxxxxxxxxxxx
OPENAI_BASE_URL=https://api.openai.com
MODEL_ID=gpt-4o

# --- Option 2: DeepSeek (Recommended for cost efficiency) ---
# If DEEPSEEK_API_KEY is set, it takes priority over OpenAI
# DEEPSEEK_API_KEY=sk-xxxxxxxxxxxxxxxx
# DEEPSEEK_BASE_URL=https://api.deepseek.com
# MODEL_ID=deepseek-chat

You can also use any OpenAI-compatible API provider, including locally deployed Ollama.

Step 3: Start the Services

docker compose up -d

This command automatically starts three services:

ServicePortDescription
frontend80React frontend application
backend8000FastAPI backend API
db5432PostgreSQL database

Step 4: Start Using

Open your browser and visit http://localhost to get started.

Try These Prompts

  • Draw a microservices architecture diagram
  • Create a user registration flowchart
  • @mindmap Organize the core concepts of React
  • @charts Create a bar chart with data: Q1: 120, Q2: 200, Q3: 150, Q4: 300

Upload Files

You can upload Excel, PDF, Word, and PPT files directly. The AI will automatically parse them and generate the corresponding charts.

Verify Service Status

# Check all container statuses
docker compose ps

# View backend logs
docker compose logs -f backend

Local Development

To develop locally with source code changes:

Backend (Python 3.13 + uv)

cd backend
uv sync                 # Install dependencies
bash start_backend.sh   # Start (includes DB migrations)

Backend runs at http://localhost:8000

Frontend (Node.js 20+)

cd frontend
npm install
npm run dev

Frontend runs at http://localhost:5173

Configure Models

After startup, you can also configure LLM models via the Settings icon in the top-right corner:

  1. Click the settings icon
  2. Enter Name, Base URL, Model ID, and API Key
  3. Select the model from the dropdown

Supports OpenAI, DeepSeek, Claude, and other mainstream models. You can also connect locally deployed models.

Next Steps