🤖 ai_code_her

Run Any AI Coding Assistant on Android — One Repo, Every Model

---

🧠 The Idea

Why install multiple apps for multiple AIs?

ai_code_her is a single lightweight Android environment that lets you connect to any AI coding assistant via Termux — no local models, no bloat, just pure API power.

Your Phone (~300MB)
    ├── Claude Code     → Anthropic API
    ├── ChatGPT         → OpenAI API
    ├── Gemini          → Google API
    ├── DeepSeek        → DeepSeek API
    └── Mistral         → Mistral API

---

✨ Features

• 🔀 Multi-AI support — switch between AIs with one command
• ⚡ One-command install — automated setup for all providers
• 🌐 Direct API connections — always latest models, no local overhead
• 🖥️ Unified server mode — expose any AI as a local HTTP endpoint
• 📁 Shared project manager — one workspace for all your AI projects
• 💾 ~300MB total — all providers combined
• 🔑 Bring your own keys — full control over your API usage

---

📋 Requirements

Requirement	Details
Device	Android 5.0+
App	Termux (from F-Droid)
Storage	~300MB free
Internet	Wi-Fi or Mobile Data
API Keys	One or more provider keys

---

🚀 Quick Install

Step 1 — Install Termux

⚠️ Use F-Droid only — not Play Store.

🔗 https://f-droid.org/packages/com.termux/

Step 2 — Run Master Bootstrap

curl -sL https://raw.githubusercontent.com/swordenkisk/ai_code_her/main/scripts/bootstrap_all.sh | bash

Step 3 — Set Your API Keys

bash ~/ai_code_her/scripts/set_keys.sh

Step 4 — Pick Your AI & Code! 🎉

ai claude     # Switch to Claude
ai chatgpt    # Switch to ChatGPT
ai gemini     # Switch to Gemini
ai deepseek   # Switch to DeepSeek
ai mistral    # Switch to Mistral

---

📂 Project Structure

ai_code_her/
├── 📄 README.md
├── 📄 LICENSE
├── 📁 claude/
│   ├── bootstrap.sh          # Claude Code setup
│   └── README.md
├── 📁 chatgpt/
│   ├── bootstrap.sh          # ChatGPT CLI setup
│   └── README.md
├── 📁 gemini/
│   ├── bootstrap.sh          # Gemini CLI setup
│   └── README.md
├── 📁 deepseek/
│   ├── bootstrap.sh          # DeepSeek setup
│   └── README.md
├── 📁 mistral/
│   ├── bootstrap.sh          # Mistral CLI setup
│   └── README.md
├── 📁 scripts/
│   ├── bootstrap_all.sh      # Install everything at once
│   ├── set_keys.sh           # API key manager
│   ├── switcher.sh           # Switch between AIs (ai command)
│   ├── server.sh             # Unified local HTTP server
│   └── new_project.sh        # Project scaffolding
├── 📁 config/
│   └── .ai_code_her.yml      # Global configuration
└── 📁 docs/
    ├── SETUP.md
    ├── KEYS.md               # Where to get each API key
    └── TROUBLESHOOT.md

---

🤖 Supported AI Providers

AI	CLI Tool	API Key Source	Size
Claude	claude-code	console.anthropic.com	~35MB
ChatGPT	chatgpt-cli	platform.openai.com	~20MB
Gemini	gemini-cli	aistudio.google.com	~20MB
DeepSeek	deepseek-cli	platform.deepseek.com	~15MB
Mistral	mistral-cli	console.mistral.ai	~15MB

---

🔀 Switching Between AIs

# Switch active AI
ai claude
ai chatgpt
ai gemini
ai deepseek
ai mistral

# Check current active AI
ai status

# Ask any AI directly
ai ask "Help me fix this bug"

# Use specific AI without switching
ai --use gemini "Explain this code"

---

🌐 Unified Server Mode

Expose any AI as a local HTTP server accessible from any device on your network:

# Start server with active AI
bash ~/ai_code_her/scripts/server.sh

# Start with specific AI
bash ~/ai_code_her/scripts/server.sh --ai gemini

# Query from any device on same Wi-Fi
curl -X POST http://YOUR_PHONE_IP:8765/ask \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Hello!", "ai": "claude"}'

---

📦 Storage Breakdown

Component	Size
Termux base	~80MB
Node.js	~55MB
Python	~15MB
Git + tools	~45MB
All AI CLIs	~105MB
Total	~300MB ✅

---

🗺️ Roadmap

• v1.1 — Add Grok (xAI) support
• v1.2 — Add Ollama bridge (local models option)
• v1.3 — AI comparison mode (same prompt → multiple AIs)
• v1.4 — Native Android GUI wrapper
• v2.0 — Full offline fallback + NetHunter bundle

---

🤝 Contributing

1. Fork the repository
2. Create your branch: git checkout -b feature/new-ai-provider
3. Commit: git commit -m "Add: new AI provider"
4. Push: git push origin feature/new-ai-provider
5. Open a Pull Request

---

📄 License

MIT License — Copyright (c) 2026 swordenkisk

See LICENSE for full details.

---

👤 Author

swordenkisk

Inventor | AI Researcher | Software Architect

🔗 GitHub: https://github.com/swordenkisk

---

"Every AI. One phone. Zero bloat." 🤖📱

⭐ Star this repo if you find it useful!