app·AI Tools

Jan

Open-source ChatGPT-style app that runs models locally on your device.

FREEOPEN SOURCELOCAL-FIRSTPRIVACY-FIRST

Setup guide · 20–35 min

Visit Jan →

Quick facts

Price model: Open source
Starting price: Free
Best for: Private chat UI · Local model management · Open-source AI desktop
Replaces: ChatGPT Plus, Poe subscription
Platforms: MacWindowsLinux
Last verified: 2026-06-22

Why it's listed

Looks like the chat apps you're used to—without uploading every prompt to someone else's server.

Jan provides a familiar chat interface while keeping conversations on your machine. It supports multiple local backends and extensions for a privacy-first alternative to cloud AI subscriptions.

The catch

Jump to setup guide ↓

Still hardware-limited; extension ecosystem is younger than cloud platforms.

How to set up Jan

ChatGPT-like desktop app that keeps conversations on your machine—open source and familiar UI.

Time: 20–35 min
Difficulty: Easy
Verified: 2026-06-22

Before you start

Mac, Windows, or Linux
16GB RAM recommended for local models
Optional: Ollama or LM Studio already running for backends

Install Jan
Download from jan.ai. Install and launch—create a local profile; cloud signup is optional.
Choose a model backend
Settings → Model Providers: Jan's bundled local runner, or connect Ollama at localhost:11434 if installed.
Download a model in Jan
Hub → pick a small chat model → download. Wait for completion before first chat.
Run a real conversation
New thread → test summarization and rewriting. Compare speed vs cloud—adjust model size.
Optional — extensions
Browse Jan extensions for extras. Disable what you don't use—keep the stack simple month one.

Troubleshooting

Cannot connect to Ollama: Ensure ollama serve running and URL is http://127.0.0.1:11434.
App hangs on generate: Model too large—switch to 3B class model.
Want team browser UI: Open WebUI on a home server (see our Open WebUI guide).