Is Jan free? Are there any hidden costs?

Jan is completely free and open source. There are no subscription fees. You only incur costs if you choose to use paid cloud APIs (like OpenAI) which you configure yourself.

How does the web search feature work if it's local?

Jan's web search feature queries search engines and fetches content to provide context to the local model. The search queries originate from your machine, but the model's reasoning stays local.

Can I use my own custom GGUF models with Jan?

Yes. Jan supports loading and running local GGUF model files. You can place them in the Jan models directory or add them through the interface.

What is Tokamak and do I need it?

Tokamak is the optional self-hosted backend for model routing, fusion, and governance. You don't need it to use Jan Desktop for local chat; it's for advanced multi-agent or enterprise setups.

Directory/Chatbots/AI Assistants/Jan

Jan

Jan is an open-source, cross-platform desktop application that lets you run large language models (LLMs) locally on your machine for complete privacy and zero API costs. It provides a clean, user-friendly interface to chat with models from OpenAI, Anthropic, Google, Meta, Mistral, Alibaba, and more, or to run open-source models like its proprietary Jan-V1-4B. Key features include a local API server, MCP connectors, and a self-hosted backend called Tokamak for model routing and governance. It's designed for developers, founders, and privacy-conscious users who want to experiment with and deploy AI assistants without vendor lock-in. The ecosystem includes Jan Desktop for personal use, Jan Agent for autonomous coding tasks, and Tokamak for enterprise-grade self-hosting.

Chatbots

Visit website

Added on

Sep 22, 2026

Local LLMsPrivacy-firstOpen sourceMulti-modelDesktop appAgent frameworkSelf-hosted backendMCP connectorsWeb search

Product information

Everything worth knowing about Jan

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Run Models Locally

Execute LLMs on your own hardware for full privacy, with no API costs or data leaving your machine.

Connect to Online Models

Plug in APIs for commercial models like ChatGPT, Claude, and Gemini from the app.

Local API Server

Jan Desktop provides a local API server to interface with the running models programmatically.

Memory Context

The system remembers your context and preferences across sessions to avoid repetition.

Technical capabilities

Jan Agent Integration

Launch autonomous coding agents against a local model via the standalone Jan Agent component.

Tokamak Router & Governance

A self-hosted backend for model switching, fusion, and audit that Jan agents connect to.

Must watch videos

Stop Sharing Your Data! Run Your Own Private AI with Jan AI

6:43

Tutorial6:43

Stop Sharing Your Data! Run Your Own Private AI with Jan AI

Tired of sending your private data to external servers? In this video, I show you how to set up your own local Large Language Model (LLM) using Jan AI — one of the easiest ways to get started with AI while keeping your data 100% on your own machine. Whether you're a researcher handling sensitive data or just curious about how models like Llama, Gemma, and DeepSeek work, this video covers everything you need to know to get up and running at home. Get smarter by subscribing to my mailing list 💻 https://www.drlyndonwalker.com [0:00] Introduction [0:26] Why use a local LLM? [0:50] Download Jan AI [1:02] Download a model [2:47] Model settings [3:17] Adding files [3:41] Testing Jan AI LLM model [4:27] Summarizing a pdf

FREE, Private & Local AI for Mac, Windows and Linux (Jan.ai)

4:54

Tutorial4:54

FREE, Private & Local AI for Mac, Windows and Linux (Jan.ai)

Learn how to run powerful AI models on Intel Macs that are local, private, and completely free. This tutorial shows you how to easily install and chat with Large Language Models (LLMs) right on your Mac, Windows, or Linux computer using Jan.AI. All while keeping your data secure and working offline. Performance depends on your computer’s specs and is often slower than cloud AI like ChatGPT, it's very usable for many tasks without any subscriptions. Watch my demo and check it out! #LocalAI #IntelMac #LLM 🔗 Links Mentioned: https://Jan.AI 💻 More Tips and Tutorials http://AnsonAlex.com 📚 LinkedIn Learning Courses https://www.linkedin.com/learning/instructors/anson-alexander 🔎 Request a Tutorial https://forms.gle/sciVnK4tHno7sCcGA ☎️ Get in Contact with Me https://www.linkedin.com/in/ansonalexander/ 🎥 Behind the Scenes https://www.instagram.com/ansonalex.c0m/

2:08

Tutorial2:08

AI Controls Your Computer with Jan AI

Jan.AI is a great local inference server for serving local LLMs. But did you know you can use it as an interface to control your computer? Jan: https://jan.ai/ Let's connect! https://x.com/MikeBirdTech https://mikebird.ca

Who uses it

Real use cases, no hype

Private AI Assistant

Chat with an AI that never sends your data to the cloud, perfect for sensitive queries and personal use.

Local Model Playground

Easily download, run, and switch between open-source LLMs like Jan-V1, Llama, and DeepSeek for experimentation.

Unified API Hub

Connect to multiple cloud providers (OpenAI, Anthropic) and use them through a single local interface.

Web-Connected AI

Perform real-time web searches and ground your AI responses in current information without leaving the app.

Developer Tooling

Use the local API server to integrate LLM capabilities directly into your own applications and workflows.

Autonomous Coding Agent

Launch Jan Agent to perform autonomous coding tasks against a local model, configured via skills and memory.

What's great

Truly private & local-first AI, no data leaves your machine
Beautiful, intuitive UI that non-engineers love
Runs open models like Jan-V1, Llama, DeepSeek locally
Supports popular cloud APIs (OpenAI, Anthropic, etc.)

Technical strengths

Open source, free, no API bills for local inference
Built-in web search and MCP connector support
Local API server for programmatic access

Where it falls short

Local model performance depends heavily on your hardware
Limited context memory for some local models
Advanced features like 'Memory' are marked 'Coming Soon'
Desktop app only, no official mobile clients

Technical limitations

Requires GGUF or MLX-compatible model files for local use
Tokamak backend adds self-hosting complexity

For technical folks

The deep-dive specs

FAQ

Includes technical Q&A

Jan runs LLM inference locally using your hardware's CPU/GPU. Your conversations and data never leave your machine unless you explicitly configure a cloud provider.