Open source alternatives

Open source ChatGPT alternatives you can run yourself

Ollama, Jan, LibreChat and Open WebUI explained: which runs models on your own computer, which gives a team one chat interface, and which licenses to read closely.

OpenSourceHunter editorsUpdated October 5, 2026

“Open source ChatGPT alternative” can mean two different things. Some projects run language models on your own hardware, so your conversations never leave your machine. Others are chat interfaces that you host yourself and connect to models, local or remote. Many setups combine one of each.

The projects below are the ones we would look at first. We checked their licenses and recent activity on October 5, 2026.

Quick comparison

Project	What it is	License	Best for
Ollama	Runs models locally behind an API	MIT	Developers and the engine for other apps
Jan	Desktop chat app for local and remote models	Apache-2.0	Private chat on one computer
LibreChat	Self-hosted, multi-user chat interface	MIT	Teams using several model providers
Open WebUI	Self-hosted chat interface	Open WebUI License	Read the branding terms first

Ollama: the local engine

Ollama downloads and runs open-weight language models on your own computer, and exposes them through a simple API. On its own it is a command-line tool, but many chat interfaces, including the ones below, can use it as their model backend.

The models themselves have their own licenses, which vary. Check each model’s terms before using it commercially.

Jan: private chat in a desktop app

Jan is a desktop app for Windows, macOS and Linux that feels like ChatGPT but runs open models offline on your computer. It can also connect to remote APIs when you want a larger model. It is licensed under Apache-2.0.

Choose it when you want private AI chat on a laptop without running a server. Speed and quality depend on your hardware, especially memory and graphics card.

LibreChat: one interface for a whole team

LibreChat is a self-hosted web app that brings many model providers into one chat interface, with user accounts, conversation search, agents, and file handling. It is MIT-licensed and very actively developed.

It is a good fit when a team wants one internal chat tool that can switch between hosted models and local ones served by Ollama. You still pay each provider for API usage.

Open WebUI: popular, with a branding condition

Open WebUI is the most widely starred interface in this space and pairs naturally with Ollama. Its license is based on the BSD-3 license but adds a condition: deployments with more than 50 users in a 30-day period may not remove or change the Open WebUI branding without permission or an enterprise license.

That condition is not part of any OSI-approved open source license, so Open WebUI is not in our directory. For many personal and small-team setups it will not matter, but read the license before you rebrand it or roll it out across a company.

How to choose

You want ChatGPT-style chat that never leaves your laptop: Jan.
You are building an app and need a local model API: Ollama.
Your team wants one internal chat interface over several providers: LibreChat, optionally backed by Ollama.
You want the Ollama-plus-web-interface setup most guides describe: Open WebUI, after reading its license.

A local model will not match the largest hosted models on every task. Try your real prompts on two or three models before deciding what is good enough. For a private place to keep what you learn, see our open source note-taking apps, or browse the directory.