Jan AI Project Detailed Overview

Project Overview

Jan is an open-source ChatGPT alternative that runs 100% offline on your computer. Developed and maintained by Menlo Research in Singapore, it aims to provide users with a completely private, self-controlled AI assistant solution.

Core Features

🏠 Local Operation

• 100% Offline Operation: Runs AI models without an internet connection, ensuring privacy and security.
• Local Data Storage: All conversations and files are stored locally, giving users full control over their data.
• No Subscription Fees: Avoids recurring payments for cloud services; download once for permanent use.

🤖 Model Support

• Local LLM Models: Supports downloading and running models like Llama, Gemma, Qwen, and others from HuggingFace.
• Cloud Integration: Can connect to cloud services such as OpenAI, Anthropic, Mistral, and Groq.
• Model Management: Provides features for browsing, downloading, importing, and configuring models.

🛠 Technical Features

• OpenAI-Compatible API: Local server runs on localhost:1337, fully compatible with the OpenAI API.
• MCP Integration: Supports Model Context Protocol, enhancing AI capabilities.
• Multi-Platform Support: Supports Windows, macOS, and Linux systems.
• GPU Acceleration: Supports NVIDIA CUDA, AMD Vulkan, and Intel Arc GPUs.

🎯 Advanced Features

• Custom Assistants: Create specialized AI assistants to perform specific tasks.
• Document Chat: Interact directly with documents like PDFs and notes to get summaries and insights.
• Extension System: Supports customizing functionalities through plugins/extensions.
• Multi-Engine Support: Compatible with various inference engines like llama.cpp and TensorRT-LLM.

System Requirements

Minimum Configuration Requirements

• macOS: 13.6+ (8GB RAM for 3B models, 16GB for 7B models, 32GB for 13B models)
• Windows: 10+, with support for NVIDIA/AMD/Intel Arc GPUs.
• Linux: Most distributions, with optional GPU acceleration.

Development Environment Requirements

• Node.js ≥ 20.0.0
• Yarn ≥ 1.22.0
• Make ≥ 3.81
• Rust (for Tauri framework)

Installation Methods

Method 1: Pre-compiled Version Download

Download the installer package for your platform from jan.ai or GitHub Releases:

Platform	Stable Version	Beta Version
Windows	jan.exe	jan.exe (nightly)
macOS	jan.dmg	jan.dmg (nightly)
Linux	jan.deb / jan.AppImage	jan.deb / jan.AppImage (nightly)

Method 2: Source Code Compilation

Build with Make

git clone https://github.com/menloresearch/jan
cd jan
make dev

Use mise Toolchain

git clone https://github.com/menloresearch/jan
cd jan
# Install mise (if not already installed)
curl https://mise.run | sh
# Install tools and start development
mise install  # Install Node.js, Rust, and other tools
mise dev      # Run full development environment

Manual Build

yarn install
yarn build:core
yarn build:extensions
yarn dev

Main Use Cases

1. Privacy Protection Scenarios

• Handling sensitive documents and confidential information.
• Environments requiring compliance.
• Avoiding data leakage to third-party servers.

2. Offline Work Environments

• Environments with limited or unstable network access.
• Air-gapped systems.
• Field operations or remote work.

3. Cost Control

• Teams with frequent AI usage.
• Avoiding usage-based cloud service fees.
• Individual users or small teams with limited budgets.

4. Customization Needs

• Scenarios requiring specific model fine-tuning.
• Custom AI assistants and workflows.
• Integration into existing applications.

Technical Architecture

Core Components

• Jan Desktop: A cross-platform desktop application based on Tauri.
• Jan Core: The core inference engine and model management.
• Jan Extensions: An extension system supporting various AI connectors and tools.
• Jan API Server: An OpenAI-compatible local API server.

Extension Ecosystem

• Cloud AI Connectors: OpenAI, Anthropic, Mistral, etc.
• Tool Integrations: Browser automation, Jupyter notebooks, code execution, etc.
• Data Connectors: Supports various data sources and formats.

Open Source License and Community

• License: Apache 2.0 Open Source License.
• Code Hosting: Publicly developed on GitHub.
• Community Support: Active Discord community and GitHub discussions.
• Contribution Guide: Community contributions are welcome; see CONTRIBUTING.md for details.

Related Links

• Official Website: https://jan.ai/
• GitHub Repository: https://github.com/menloresearch/jan
• Documentation Center: https://jan.ai/docs
• API Reference: https://jan.ai/api-reference
• Discord Community: https://discord.gg/AsJ8krTT3N
• Changelog: https://jan.ai/changelog

Summary

Jan AI is a powerful, privacy-focused open-source AI assistant platform that provides users with a complete solution for running large language models locally. Whether for individual users or enterprise teams, Jan offers a highly controllable, secure, and private AI experience, while avoiding the costs and privacy concerns associated with cloud services.