README.md

metadata

title: Axon v6
emoji: 🔥
colorFrom: indigo
colorTo: blue
sdk: docker
pinned: true
license: mit
short_description: AI Coding Assistant - 8 Models - 19 Tools - 100% Local

🔥 Axon v6

The Ultimate Free AI Coding Assistant

---

✨ Features

🤖 8 Powerful Models

Model	Size	Best For
⭐ Qwen3 Coder 30B-A3B	~10GB	Best quality (MoE)
🏆 Qwen2.5 Coder 14B	~8GB	Premium tasks
🧠 DeepSeek V2 Lite	~9GB	Complex logic
⚖️ Qwen2.5 Coder 7B	~4.5GB	Balanced
🚀 Qwen2.5 Coder 3B	~2GB	Fast & capable
⚡ DeepSeek Coder 6.7B	~4GB	Algorithms
💨 Qwen2.5 Coder 1.5B	~1GB	Quick tasks
🔬 Qwen2.5 Coder 0.5B	~0.3GB	Instant

🛠️ 19 Tools

Core: Chat, Generate, Explain, Debug, Review

Advanced: Security Scan, Complexity Analysis, Convert, Test, Document, Optimize, Diff, Pseudocode, Interview

Builders: SQL, Shell, Cron, Regex, API

Data: Mock Data Generator, Format Converter

🎤 Voice Input

Whisper-powered speech-to-text - just speak your code requests!

💾 Export

Save your chat history and generated code

---

🚀 Tech Stack

• Inference: llama.cpp via llama-cpp-python
• Wheels: AIencoder/llama-cpp-wheels (pre-built for Debian/Ubuntu)
• UI: Gradio
• Speech: faster-whisper
• Models: GGUF format from HuggingFace

---

💡 Why Axon?

Feature	Axon	Others
100% Local	✅	❌
No API Keys	✅	❌
No Rate Limits	✅	❌
Free Forever	✅	❌
Privacy	✅	❌
8 Models	✅	❌
19 Tools	✅	❌

---

🛞 Pre-built Wheels

pip install https://huggingface.co/datasets/AIencoder/llama-cpp-wheels/resolve/main/llama_cpp_python-0.3.16-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl

---

🙏 Credits

• Qwen for amazing coding models
• DeepSeek for logic-focused models
• ggerganov for llama.cpp
• Gradio for the UI framework

---

Built with ❤️ by AIencoder