VersaAI 🤖

Production-Grade Multi-Model AI Platform for Developers

VersaAI is an intelligent code assistant that combines multiple AI models (local and cloud) to provide powerful coding assistance through a CLI, code editor integration, and Python API. Built for privacy, performance, and flexibility.

✨ Features

🤖 Multi-Model AI - Use DeepSeek, StarCoder, CodeLlama, Qwen, GPT-4, Claude
🧠 Intelligent Routing - Automatically selects best model for each task
💬 CLI Code Assistant - Interactive terminal-based AI pair programmer
🔌 Editor Integration - Real-time AI assistance in NLPL Code Editor
📚 RAG System - Understands your codebase with semantic search
🔒 Privacy-First - Local models keep your code 100% private
⚡ High Performance - Fast responses with optimized model loading
💰 Cost-Effective - Free local models + optional cloud APIs

🚀 Quick Start

Option 1: Full Stack (CLI + GUI + Editor Integration)

./launch.sh
# Select: 1. Full Stack (Backend + Flutter UI)

Option 2: CLI Code Assistant Only

versaai
# Or: python3 versaai_cli.py

Option 3: Flutter Desktop UI

cd ui && ./scripts/run_with_backend.sh

Option 4: Code Editor Integration

python3 start_editor_bridge.py  # Backend server
# Then open NLPL Code Editor - AI features auto-connect

Choose Your Model:

Local Models (Free, Private) - DeepSeek, StarCoder, CodeLlama, WizardCoder
OpenAI API - GPT-4, GPT-3.5-Turbo (requires API key)
Anthropic API - Claude 3.5, Claude 3 (requires API key)

📖 See INTEGRATION_COMPLETE.md for full documentation

💬 Chat Mode:
> Explain async/await in Python

📝 Code Mode:
> /code Write a function to validate emails

🔧 File Mode:
> /file mycode.py  # Analyze specific file

That's it! You're now coding with AI assistance.

📖 Documentation

📘 Complete User Guide - Everything you need to know
🚀 Quick Start Guide - Get started in 5 minutes
📚 User Manual - In-depth documentation
🎓 Tutorials - Step-by-step guides
🔌 Editor Integration - Integrate with your editor
🔀 Multi-Model Setup - Use multiple models
⚡ Quick Reference - Command cheat sheet

🎯 Use Cases

Use Case	Description
Code Completion	Real-time suggestions as you type
Code Explanation	Understand unfamiliar code
Refactoring	Improve code quality and structure
Debugging	Find and fix bugs with AI help
Test Generation	Auto-generate unit tests
Documentation	Add comments and docstrings
Code Review	Get AI feedback on your code
Learning	Ask questions about programming concepts

🔧 Installation Options

Option 1: Basic (CLI Only)

pip install -e .
versaai

Option 2: Full (CLI + Editor + All Features)

pip install -e ".[all]"

# Download models
versaai  # Interactive model downloader

# Start editor backend
python -m versaai.code_editor_bridge.server

Option 3: Docker

docker pull versaai/versaai:latest
docker run -it versaai/versaai

🧠 Available Models

Local Models (Free, Private)

Model	Size	RAM	Best For
DeepSeek Coder 1.3B	834MB	2GB	Fast completions
DeepSeek Coder 6.7B	4.1GB	8GB	Balanced quality
StarCoder2 7B	4.3GB	8GB	Code generation
CodeLlama 7B	4.0GB	8GB	General coding
Qwen2.5-Coder 7B	4.4GB	8GB	Latest, high quality

Cloud Models (Paid, Powerful)

Model	Provider	Best For
GPT-4	OpenAI	Complex tasks, architecture
GPT-3.5 Turbo	OpenAI	Fast, cost-effective
Claude 3 Sonnet	Anthropic	Explanations, docs
Claude 3 Opus	Anthropic	Highest quality

🔌 Code Editor Integration

VersaAI integrates with the NLPL Code Editor for real-time AI assistance.

Setup

# Terminal 1: Start VersaAI backend
cd VersaAI
python -m versaai.code_editor_bridge.server

# Terminal 2: Start Code Editor
cd code_editor
npm install
npm run dev

Features in Editor

AI Chat Panel - Ctrl+Alt+V
Code Completions - Automatic suggestions
Explain Code - Right-click → Explain
Refactor - Right-click → Refactor
Generate Code - Right-click → Generate
Debug - Right-click → Debug

See Editor Integration Guide for details.

🎓 Examples

Example 1: Code Explanation

$ versaai

> Explain this code:
> def fibonacci(n):
>     return n if n <= 1 else fibonacci(n-1) + fibonacci(n-2)

AI: This is a recursive implementation of the Fibonacci sequence.
    It returns n for base cases (0 or 1), otherwise recursively
    calculates fib(n-1) + fib(n-2). Note: This has exponential
    time complexity O(2^n). Consider using memoization or iteration
    for better performance.

Example 2: Code Generation

> /code Write a function to validate email addresses with regex

AI: Here's a robust email validation function:

    import re
    
    def validate_email(email: str) -> bool:
        """
        Validates email address format.
        
        Args:
            email: Email address to validate
            
        Returns:
            True if valid, False otherwise
        """
        pattern = r'^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$'
        return bool(re.match(pattern, email))

Example 3: Multi-Model Mode

$ versaai --multi-model

> Complete: def quicksort(arr):
AI: [Using deepseek-1.3b for fast completion]
    def quicksort(arr):
        if len(arr) <= 1:
            return arr
        pivot = arr[len(arr) // 2]
        left = [x for x in arr if x < pivot]
        middle = [x for x in arr if x == pivot]
        right = [x for x in arr if x > pivot]
        return quicksort(left) + middle + quicksort(right)

> Design a scalable microservices architecture
AI: [Routing to GPT-4 for complex architecture]
    Here's a comprehensive microservices architecture...

🐛 Troubleshooting

Issue: "Model not found"

# Download model
versaai  # Select "Download new model"

# Or manual download
cd ~/.versaai/models
wget <model-url>

Issue: "Out of memory"

# Use smaller model
versaai --model deepseek-1.3b

# Or reduce context window
export VERSAAI_MAX_TOKENS=512

Issue: "Can't connect to editor backend"

# Check if backend is running
ps aux | grep code_editor_bridge

# Start it
python -m versaai.code_editor_bridge.server

See Troubleshooting Guide for more.

🔒 Privacy & Security

Local Models

✅ 100% offline, no data leaves your machine
✅ No telemetry, tracking, or analytics
✅ Your code stays private

Cloud APIs (Optional)

⚠️ Data sent to OpenAI/Anthropic servers
⚠️ Subject to provider's privacy policy
✅ Can be disabled with VERSAAI_LOCAL_ONLY=true

📊 Benchmarks

Task	DeepSeek 1.3B	DeepSeek 6.7B	GPT-4
Code Completion	2s	5s	3s
Code Generation	3s	8s	5s
Explanation	4s	10s	6s
Quality (1-10)	7/10	8.5/10	9.5/10
Cost	Free	Free	~$0.03/request

🤝 Contributing

Contributions welcome! See CONTRIBUTING.md for guidelines.

Development Setup

# Clone
git clone https://github.com/The-No-hands-Company/VersaAI.git
cd VersaAI

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black versaai/
isort versaai/

📄 License

MIT License - See LICENSE for details.

🙏 Acknowledgments

Built with:

llama.cpp - Fast local model inference
OpenAI & Anthropic - Cloud API support
Hugging Face - Model hosting
langchain - RAG framework

📞 Support

📖 Documentation: See guides above
🐛 Bug Reports: Open an issue on GitHub
💡 Feature Requests: Open an issue with [Feature] tag
💬 Questions: Open a discussion on GitHub

Made with ❤️ by The No-hands Company

Start coding smarter with AI today! 🚀

pip install -e .
versaai

Name		Name	Last commit message	Last commit date
Latest commit History 16 Commits
.github		.github
.vscode		.vscode
bindings		bindings
config		config
desktop		desktop
docs		docs
examples		examples
include		include
plugins		plugins
scripts		scripts
src		src
tests		tests
ui		ui
versaai		versaai
.clang-format		.clang-format
.clang-tidy		.clang-tidy
.cursorrules		.cursorrules
.editorconfig		.editorconfig
.git-blame-ignore-revs		.git-blame-ignore-revs
.gitignore		.gitignore
.gitmodules		.gitmodules
.pre-commit-config.yaml		.pre-commit-config.yaml
AUTHORS.md		AUTHORS.md
CHANGELOG.md		CHANGELOG.md
CMakeLists.txt		CMakeLists.txt
CODEOWNERS		CODEOWNERS
EULA.txt		EULA.txt
README.md		README.md
debug_imports.py		debug_imports.py
download_model.py		download_model.py
launch.sh		launch.sh
pyproject.toml		pyproject.toml
start_editor_bridge.py		start_editor_bridge.py
test_coding_agent_cli.py		test_coding_agent_cli.py
test_mandate.cpp		test_mandate.cpp
test_mandate.sh		test_mandate.sh
test_mandate_demo.sh		test_mandate_demo.sh
uv.lock		uv.lock
verify_agents.py		verify_agents.py
verify_chat_integration.py		verify_chat_integration.py
verify_code_model.py		verify_code_model.py
verify_setup.py		verify_setup.py
versaai_cli.py		versaai_cli.py

Folders and files

Latest commit

History

Repository files navigation