Skip to content

AchuAshwath/drawmaid

Drawmaid

Voice and text to Excalidraw diagrams using local AI. Generate beautiful diagrams from natural language — entirely in your browser or connect to your own AI server.

Features

  • Voice-to-Diagram — Speak your diagram ideas and watch them appear
  • Text-to-Diagram — Type natural language and get instant Mermaid diagrams
  • On-Device AI — WebLLM runs entirely in your browser, no server needed
  • Local Server Support — Connect to OpenCode, Ollama, LM Studio, or any OpenAI-compatible API
  • Works Offline — Once downloaded, WebLLM works without internet
  • Excalidraw Integration — Full Excalidraw editing capabilities after generation

Quick Start

Option 1: WebLLM (No Setup Required)

  1. Visit https://app.drawmaid.ashwath.space/
  2. Download a WebLLM model (recommended: Qwen2.5-Coder-1.5B)
  3. Start generating diagrams!

Option 2: Local Server (More Powerful)

  1. Install OpenCode
  2. Run opencode serve --cors https://drawmaid.ashwath.space in your terminal, if you're trying to connect to https://drawmaid.ashwath.space
  3. Run opencode serve if you're developing from localhost
  4. In Drawmaid, go to AI Configuration → Local Server → Connect

Installation (Development)

# Clone the repository
git clone https://github.com/AchuAshwath/drawmaid.git
cd drawmaid

# Install dependencies
bun install

# Start development server
bun dev

The app will be available at http://localhost:5173

AI Configuration

WebLLM (Browser-based)

WebLLM runs AI models directly in your browser using WebGPU. No data leaves your device.

  • Pros: Privacy-friendly, works offline, no server setup
  • Cons: Requires WebGPU-capable browser, model downloads to device

Recommended models:

  • Qwen2.5-Coder-1.5B-Instruct-q4f16_1-MLC — Best for diagrams

Local Server Options

Provider Description Default URL Status
OpenCode Full-featured local AI server http://127.0.0.1:4096 ✅ Available
Ollama Run Llama, Mistral, and other models locally http://localhost:11434/v1 Coming Soon
LM Studio Desktop app for running LLMs http://localhost:1234/v1 Coming Soon
vLLM High-performance LLM serving http://localhost:8000/v1 Coming Soon

Tech Stack

  • Frontend: React 19, TanStack Router, Vite, Tailwind CSS
  • AI: WebLLM (on-device), OpenAI-compatible APIs
  • Diagrams: Excalidraw, Mermaid
  • Testing: Vitest, Playwright

License

AGPL-3.0 — See the LICENSE file for details.

Contributing

Contributions are welcome! Please read our contributing guidelines before submitting PRs.


Built with ❤️ using React Starter Kit, WebLLM and Excalidraw

About

Voice assisted excali draw companion that can visualize diagrams for you as you speak. Open source project that helps you not get yet another AI subscription and reuse

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Sponsor this project

  •  

Packages

 
 
 

Contributors