Does Jan Support Multiple LLM Backends On The Same Machine?

Yes. Jan supports llama.cpp (CPU/GPU), Apple MLX for Apple Silicon, AMD ROCm/HIP on Linux, and NVIDIA CUDA — all selectable per model, letting you mix inference backends based on your hardware configuration.

Can Jan Be Used As A Drop-In Replacement For The OpenAI API?

Jan exposes a fully OpenAI-compatible REST API at http://127.0.0.1:1337, meaning any application or library that targets the OpenAI API can point to Jan's local server instead — no code changes required.

What Is MCP Support In Jan And What Does It Enable?

MCP (Model Context Protocol) in Jan lets the AI interact with external tools and services — including browser control, web search, and integrations like Linear, Todoist, and Jupyter — enabling true agentic, multi-step task execution.

How Does Jan Handle Model Context Length And RAM Usage?

Jan v0.7.9 and later automatically cap context length during model loading to prevent excessive RAM consumption. Per-model fit labels and auto-tuned llama.cpp settings help match model size to your available hardware.

What Is The Difference Between Jan's Local And Cloud Model Modes?

In local mode, all inference runs on your machine with zero network calls — your conversations never leave your device. In cloud mode, Jan acts as a unified interface connecting to providers like OpenAI, Anthropic, Groq, and OpenRouter using your own API keys.

Listed · Reviewed Jul 2026 · 7-criteria rubric

AI Productivity Tools

Jan

Name: Jan
Availability: InStock
Author: Menlo Research

Run Open-Source AI Locally With Complete Privacy And Control

by Menlo Research · Ho Chi Minh City, Vietnam · Founded 2023

Visit Jan →

Pricing

100% Free & Open Source (Apache 2.0)

What is Jan?

Jan is a free, open-source desktop application that turns your computer into a fully private AI workstation. Built by Menlo Research, it lets you download and run large language models — from Llama, Mistral, Qwen, and Gemma — entirely offline on your own hardware.

It also connects to cloud providers like OpenAI, Anthropic, and Groq when needed.

With a built-in OpenAI-compatible local API server, MCP (Model Context Protocol) support for agentic tasks, and zero data collection by design, Jan is the go-to choice for privacy-first AI users, developers, and anyone looking to break free from subscription-based cloud AI.

Jan — Run Open-Source AI Locally With Complete Privacy And Control Whether you're evaluating Jan for your team or comparing it to alternatives in the AI Productivity Tools category, this in-depth review covers everything: features, pricing, real user reviews, pros and cons, integrations, and direct comparisons against competitors.

Key Features 8

100% Offline Local AI That Runs Open-Source LLMs with Full Privacy

Supports Llama, Gemma, Qwen, DeepSeek, Mistral, and GPT-oss Models

Built-In OpenAI-Compatible Local API Server at localhost:1337

Connect to Cloud Models Like ChatGPT, Claude, Gemini, and Groq

Model Context Protocol (MCP) Support for Agentic and Tool-Calling Workflows

Create Custom AI Assistants with Personalized System Prompts and Configurations

Cross-Platform Desktop App for macOS, Windows, and Linux with GPU Acceleration

Persistent Memory Feature to Retain User Context and Preferences Across Sessions

Who Is Jan For

1 Privacy-Conscious Professionals

2 Software Developers And Engineers

3 AI Researchers

4 Self-Hosted AI Enthusiasts

5 Data Scientists

6 Open-Source Contributors

Pros & Cons

Pros 4 benefits

Zero Data Leaves Device
43.5K+ GitHub Stars
True Multi-Backend GPU Support
Active Weekly Release Cadence

Cons 3 limitations

Heavy Hardware Requirements
No Mobile App Support
Cloud UI Lags Behind

Frequently Asked Questions

5 questions

Who is Jan for?

Jan is most useful for Privacy-Conscious Professionals, Software Developers And Engineers, AI Researchers and Self-Hosted AI Enthusiasts.

Jan pricing

Jan is free to use. 100% Free & Open Source (Apache 2.0). For the current tier breakdown and any limits, see the pricing section above or check the vendor's pricing page directly — limits and prices change.

What's New

weekly

Message Branching, HTML/SVG Artifacts & Video Input 0.8.3

Introduces navigable message version branching, interactive HTML and static SVG artifact previews, video input for local vision models, a unified reasoning/tool chain-of-thought timeline, and a fully custom Linux titlebar.

Jun 24

AMD ROCm/HIP On Linux & Resumable Downloads 0.8.2

Adds AMD ROCm/HIP backend support on Linux, phased-startup loader for faster app launch, pause/resume model downloads, and a safer default context size.

Jun 1

View all updates

User Base

6M+ Downloads

Active Users

Security & Privacy

Local Device Only

Apache 2.0 Open Source License Local-Only Data Storage (No PII Collected)

All Conversation Data Stored Locally On Device No Telemetry Or Personal Data Transmitted By Default Fully Auditable Open-Source Codebase Granular Privacy Controls In App Settings Optional Analytics Opt-Out Available

Collaboration & Teams

Version History

Learning & Support

Resources

Documentation Blog

Community

Discord

Support Channels

Email Onboarding

Localization

UI Languages

Content Languages

Recognition & Trust

Featured on PH Open Source

Media: Featured on GitHub Trending, covered by It's FOSS, OpenReplay Blog, and Petronella Tech

All Features of Jan