🧠 CodeThinker

AI coding assistant with chain-of-thought, 6 modes, WebLLM local inference & multi-model fallback.

✨ Features

• 6 Thinking Modes — vibe · debug · refactor · architect · explain · search
• Chain-of-Thought Streaming — watch the AI reason step-by-step in real time
• Local AI (WebLLM) — run Qwen2.5-1.5B fully in-browser via WebGPU, zero data leaves your machine
• Multi-Model Fallback — Groq (llama-3.3-70b) → Groq fallbacks → DeepSeek V3 → optional GPT-4o
• Real-Time Tool Use — web_search (Tavily), crypto_price (CoinGecko), get_datetime with parallel execution
• Voice Input & Vision — paste screenshots, speak your prompts
• Chat Persistence — sessions saved to Supabase, full history in UI

🚀 Quick Start

git clone https://github.com/GodLocal2026/codethinker.git
cd codethinker
npm install
cp .env.example .env.local
# Fill in your API keys in .env.local
npm run dev

Open http://localhost:3000

🏗️ Architecture

codethinker/
├── app/                    # Next.js 14 App Router
│   ├── api/
│   │   ├── chat/           # Main inference endpoint
│   │   └── tools/          # Tool-use handlers
│   └── page.tsx            # Main UI
├── components/
│   ├── CodeThinker.tsx     # Core chat component
│   ├── ModeSelector.tsx    # 6-mode switcher
│   ├── ThinkingSteps.tsx   # CoT streaming display
│   └── LocalAIToggle.tsx   # CLOUD/LOCAL switch
├── lib/
│   ├── providers/          # AI provider clients
│   ├── tools/              # Tool definitions
│   └── webllm/             # WebLLM/WebGPU integration
└── types/

🔧 Environment Variables

# Required
GROQ_API_KEY=your_groq_key

# Optional (for tool-use web search)
TAVILY_API_KEY=your_tavily_key

# Optional (for chat persistence)
NEXT_PUBLIC_SUPABASE_URL=your_supabase_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_key

# Optional (users can provide their own key in Settings)
OPENAI_API_KEY=

# Optional (emergency fallback)
DEEPSEEK_API_KEY=

🤖 AI Models

Priority	Provider	Model	Notes
1	Groq	llama-3.3-70b-versatile	Primary
2	Groq	llama-3.1-8b-instant	Fallback
3	Groq	gemma2-9b-it	Fallback
4	DeepSeek	deepseek-chat (V3)	Emergency fallback
5	OpenAI	gpt-4o	User-provided key
6	WebLLM	Qwen2.5-1.5B	100% local, WebGPU

🛠️ Modes

Mode	Icon	Purpose
vibe	✨	Creative coding, ideas
debug	🐛	Find and fix bugs
refactor	🔧	Clean up and optimize code
architect	🏛️	System design, structure
explain	📖	Understand code
search	🌐	Web search + code answers

🤝 Contributing

PRs welcome! See CONTRIBUTING.md for guidelines.

1. Fork the repo
2. Create your branch (git checkout -b feat/amazing-feature)
3. Commit your changes (git commit -m 'feat: add amazing feature')
4. Push to the branch (git push origin feat/amazing-feature)
5. Open a Pull Request

📄 License

MIT © GodLocal

---

🌐 Live Demo · 🐛 Report Bug · 💡 Request Feature