Xendi AI - Autonomous AI Coding Assistant

Your intelligent coding companion powered by local LLMs

Features • Installation • Usage • Configuration • Support

🚀 Features

🤖 Autonomous Agent Mode

Works like GitHub Copilot - explores, analyzes, and fixes code autonomously
Multi-step task execution without manual intervention
Automatically reads files, finds bugs, and implements fixes
Tests changes and iterates until task completion

💬 Interactive Chat Mode

Ask questions about your code
Get explanations and suggestions
Code generation and refactoring assistance

🌍 Universal Language Support

Supports all programming languages:

JavaScript/TypeScript, Python, Java, C++, C#, Go, Rust
PHP, Ruby, Swift, Kotlin, Dart, and more

🛠️ Powerful Tools

File Operations: Read, write, search, and modify files
Code Analysis: Search patterns, find bugs, analyze structure
Shell Commands: Run tests, builds, install packages
Web Search: Google search and web content fetching
Memory: Save and recall information across sessions

🔒 Privacy First

Runs 100% locally using LM Studio
Your code never leaves your machine
No telemetry, no tracking, no cloud services

🎨 Multilingual Interface

English and German UI
Easy language switching in settings

📦 Installation

Prerequisites

Visual Studio Code (v1.83.0 or higher)
LM Studio - Download from lmstudio.ai
A compatible LLM model (recommended: Qwen 2.5 Coder, DeepSeek Coder, or Llama 3.2)

Install Extension

Open VS Code
Go to Extensions (Ctrl+Shift+X / Cmd+Shift+X)
Search for "Xendi AI"
Click Install

Setup LM Studio

Download and install LM Studio
Download a model (recommended: Qwen/Qwen2.5-Coder-7B-Instruct)
Start the local server (default: http://localhost:1234)
Keep LM Studio running while using Xendi AI

🎯 Usage

Quick Start

Open Xendi AI: Click the Xendi AI icon in the Activity Bar (left sidebar)
Select Model: Choose your loaded model from the dropdown
Choose Mode:
- Chat Mode: Ask questions and get answers
- Agent Mode: Let AI autonomously complete tasks

Agent Mode Examples

Fix bugs autonomously:

Fix the TypeScript compilation errors

Create new features:

Create a REST API with Express.js including CRUD operations for users

Refactor code:

Refactor the authentication module to use async/await

Debug issues:

The login is not working, find and fix the issue

Chat Mode Examples

Code explanation:

Explain what this function does

Code generation:

Write a function to validate email addresses

Debugging help:

Why am I getting a null pointer exception here?

⚙️ Configuration

Settings

Access settings via the Settings tab in Xendi AI sidebar:

Server URL: LM Studio server address (default: http://localhost:1234)
Language: Interface language (English/German)
Auto-approve Tools: Enable automatic execution of specific tools in Agent mode

Recommended Model Settings (in LM Studio)

Temperature: 0.1 - 0.3 (lower = more deterministic)
Max Tokens: 2000+
Context Length: 4096+ (higher = better for large files)

Best Models for Xendi AI

Model	Size	Best For
Qwen 2.5 Coder 7B	7B	General coding, debugging
DeepSeek Coder 6.7B	6.7B	Code generation, analysis
Llama 3.2 3B	3B	Quick responses, chat
CodeLlama 13B	13B	Complex refactoring (needs GPU)

📚 How It Works

Agent Mode Workflow

Explore: Lists directories, finds relevant files
Analyze: Reads code, understands structure
Plan: Identifies what needs to be done
Execute: Makes changes to files
Verify: Runs tests and checks for errors
Iterate: Repeats until task is complete

🔧 Available Tools

Tool	Description	Example
`listDirectory`	List directory contents	Explore project structure
`glob`	Find files by pattern	Find all TypeScript files
`readFile`	Read file contents	Analyze code
`writeFile`	Create/update files	Generate new code
`replace`	Replace text in files	Fix bugs, refactor
`runShellCommand`	Execute commands	Run tests, build
`searchFileContent`	Search across files	Find specific code
`readManyFiles`	Read multiple files	Analyze related code

🐛 Troubleshooting

Models not loading?

Ensure LM Studio is running on http://localhost:1234
Check the server URL in Xendi AI Settings
Verify a model is loaded in LM Studio

Agent not working correctly?

Use a capable model (Qwen 2.5 Coder or DeepSeek Coder recommended)
Set temperature to 0.1-0.3 in LM Studio
Check Logs tab for error messages

Commands not executing?

Enable auto-approve for tools in Settings
Check workspace permissions
Ensure commands are valid for your OS

📝 License

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

👨‍💻 Author & Support

Developer: Robin Oliver Lucas Publisher: RL-Dev Website: https://rl-dev.de

Support & Feedback

🐛 Report bugs: GitHub Issues
💡 Feature requests: GitHub Discussions
📧 Contact: https://rl-dev.de

🙏 Acknowledgments

Built with LM Studio API
Inspired by GitHub Copilot and Claude
Community feedback and contributions

Made with ❤️ by RL-Dev