apex-coder 1.0.0

APEX - AI powered terminal coding assistant

⚡ APEX — Your AI Coding Assistant

APEX is a terminal-based agentic AI coding assistant that helps you write code, fix bugs, run commands, search your project, manage GitHub repos, and automate tasks — all from your terminal.

---

📸 Preview

Dashboard	Auto Mode	Code Generation

GitHub	RAG Search	Analyze & Test

---

✨ Features

• ⚡ Fast AI Responses — Powered by Groq (llama-3.1, llama-3.3) for ultra-fast inference
• 🧠 Smart Automation — /auto on lets APEX write, run, and fix code automatically
• 📁 RAG — Project Search — Index your codebase and search it with natural language
• 🔧 Tool Execution — Run terminal commands safely with dangerous command detection
• 🐙 GitHub Integration — List, search, push, pull, create repos from the terminal
• 👀 Watchdog — Monitor files and auto-analyze on every save
• 🔍 Code Analysis — Scan files for real bugs and auto-fix them
• 🧪 Unit Test Generator — Auto-generate and run tests in a safe sandbox
• 💾 Save Code — Save any code block from responses with /save
• 🌀 Mistral Fallback — Automatically falls back to Mistral if Groq is unavailable

---

🚀 Installation

pip install apex-coder

Or clone and install manually:

git clone https://github.com/parinprajapati/apex.git
cd apex
pip install -e .

---

🔑 Setup

On first run, APEX will ask for:

• Groq API key — get it from https://console.groq.com
• Mistral API key — get it from https://console.mistral.ai
• Your experience level (Beginner / Intermediate / Advanced)
• Whether to allow auto command execution

python -m apex.agent.main

---

🛠️ Commands

+---------------------------+---------------------------------------------------+

Command	Description
/index <path>	Index a project folder for RAG code search
/save <file>	Save the last code block to a file
/github	GitHub actions (login, push, pull, search, etc.)
/watch <file>	Monitor a file and auto-analyze on every save
/analyze <file>	Scan a file for bugs and bad practices
/fix <file>	Auto-fix issues in a file (shows preview first)
/test <file>	Generate and run unit tests in a safe sandbox
/auto on|off	Toggle smart automation mode
/setting	View or update your config (API keys, level)
/help	Show full help guide
exit	Quit APEX
+---------------------------+---------------------------------------------------+

---

🤖 Auto Mode

Enable auto mode and just describe what you want:

>: /auto on
>: create a hello world python script and run it

APEX will write the file, run it, and show you the output — fully automated.

---

📁 RAG — Project Code Search

>: /index C:\Users\Parin\Projects\MyApp
>: find the login function
>: where is the database connection?

Supports: .py .js .ts .java .cpp .c .cs .kt .scala .html .md

---

🐙 GitHub Commands

/github login <token>
/github list
/github search <keyword>
/github commits <owner/repo>
/github make <repo-name>
/github push <owner/repo> <file> <message>
/github pull <owner/repo> <file>
/github view <owner/repo> <file>

---

👀 Watchdog

/watch myfile.py       # auto-analyze on every save
/analyze myfile.py     # one-time scan
/fix myfile.py         # auto-fix with preview
/test myfile.py        # generate + run tests
/wdstop                # stop watching

---

🧠 Models Used

+-------------------+---------------------------------------------+

Purpose	Models
General coding	llama-3.1-8b-instant, llama-3.3-70b
Automation	mistral-large-latest, codestral-latest
Router / Planner	mistral-small-latest, ministral-3b
Watchdog / Fix	llama-3.3-70b, codestral-latest
+-------------------+---------------------------------------------+

---

📂 Project Structure

apex/
├── agent/          # Core agent loop, planner, UI
├── auto/           # Auto mode pipelines, stages, memory
├── config/         # Settings, system prompt, help
├── github/         # GitHub integration
├── models/         # LLM fallback + model routing
├── rag/            # RAG indexing and code search
├── tools/          # File, command, and code tools
└── watchdog/       # File watcher, analyzer, fixer, tester

---

🔒 Safety

APEX detects and blocks dangerous commands like rm -rf, format, drop table, shutdown, and many more across Windows, Linux, and macOS. You can also set auto_run_commands: false to approve every command manually.

---

👨‍💻 Author

Parin Prajapati

• 📧 [e-mail]
• 🐙 GitHub
• 🎂 Age: 19

---

📄 License

This project is licensed under the MIT License.

---

⭐ Support

If you find APEX useful, please give it a ⭐ on GitHub — it means a lot!