Draw Agent Studio

Draw and edit through conversation.

简体中文 · English

Demos

Draw a seaside lighthouse through conversation

language-demo.mp4

Create a research flowchart through conversation

research-demo.mp4

Finish an AI sunflower sketch by hand

drawing-demo.mp4

Features

• Describe: Create sketches, illustrations, and flowcharts from text.
• Refine: Move objects, change colors, delete, undo, and redo through conversation.
• Draw by hand: Select, drag, use the pen, or add text directly on the canvas.
• Save your work: Canvases save automatically. Export PNG, JPG, WebP, SVG, or editable Excalidraw files.

Quick start

Requires Python 3.11+ and Node.js 20.19+. On Windows, install Chrome as well.

1. Install and start

git clone https://github.com/bugkeep/ai-draw.git
cd ai-draw
python -m pip install -r requirements.txt
cd frontend-app
npm ci
npm run build
npx playwright install --with-deps chromium
cd ..
python run.py

For later launches, run python run.py from the project directory.

2. Connect a model

Open http://127.0.0.1:8000 and click Settings (设置) in the top-right corner.

1. Choose a provider and enter the model name, API URL, and API key.
2. Use a model with image input and tool calling, then click Test connection (测试连接).
3. Once the test passes, click Save (保存配置) and return to the canvas.

Choose OpenAI, Claude, Gemini, Ark, or another provider, or connect a local service or custom compatible endpoint. API usage is billed by your provider. Model setup

Try these prompts

• "Draw a multimodal RAG research flowchart, grouped into data preparation, retrieval and generation, and evaluation and output."
• "Add a quality check and a feedback loop to the evaluation section."
• "Draw a seaside lighthouse under moonlight, with a small sailboat and stars."
• "Make the sea a deeper blue-green and brighten the lighthouse light."

Feedback and contributions

Share your drawings, suggest ideas, or report problems in Issues.

Deployment · Contributing · Development · MIT License