The open-source alternatives to ChatGPT are Open WebUI, LibreChat, AnythingLLM, Jan, LobeHub, Chatbox, Cherry Studio, TextGen, Chat UI and Khoj. For a team on its own server, Open WebUI and LibreChat bring documents, web search, roles and sign-in. For one person, Jan runs open models on a laptop with nothing sent out, AnythingLLM chats with your documents with citations, and Chatbox covers desktop and phone. Two, Open WebUI and LobeHub, are source available. A team switches to choose its own model and keep conversations private.
8 of 10 carry an OSI licence. 10 of 10 run on your own machines. Figures read from GitHub on September 17, 2026.
One project per need, with the reason from its own README and the one thing to know before you choose it.
Documents over 9 vector databases, web search, image generation, voice, and roles with LDAP, OAuth single sign-on and SCIM, on your own server with Ollama or any OpenAI-compatible model.
Know firstSource available: above 50 users you may not remove its branding without permission or a paid licence.
Anthropic, OpenAI, Google, Bedrock, Azure and Ollama in one interface, with agents, MCP, a code sandbox for eight languages, web search, artifacts and sign-in.
Know firstA server to run with Docker, with no desktop or mobile app, and every GitHub release tagged pre-release.
MIT, with document chat and source citations, agents with MCP, and 39 model providers including local llama.cpp and Ollama, as a desktop app or a Docker server.
Know firstMulti-user support is Docker only, and telemetry is on until you turn it off.
Apache-2.0 desktop app that runs open models from Hugging Face on your computer, connects to hosted models when you choose, and serves a local OpenAI-compatible API.
Know firstNo document chat, web search or image generation described, and local quality depends on your RAM.
GPL-3.0 client for OpenAI, Claude, Gemini and local Ollama models on Windows, macOS, Linux, iOS and Android, with conversations stored on the device.
Know firstNo document chat, web search or MCP described; a separate pro edition exists.
AGPL-3.0 client with OpenAI, Gemini, Anthropic, Ollama and LM Studio, 300+ preset assistants, side-by-side model chats, file support and MCP.
Know firstSpeech, OCR and mobile apps are still on its roadmap.
Agent builder, agent groups, scheduled runs, editable memory and a library of 10,000+ tools and MCP plugins.
Know firstSource available: distributing a modified version needs a paid licence, and the README names only OpenAI in setup.
AGPL-3.0, five engines including llama.cpp and ExLlamaV3, an OpenAI and Anthropic-compatible API, tools, file chat, image generation and LoRA training, fully offline.
Know firstBuilt for technical users; non-GGUF formats need a 10 GB full install.
Apache-2.0 SvelteKit app that powers HuggingChat, for any OpenAI-compatible endpoint, with MongoDB storage, MCP tools and a per-message model router.
Know firstOpenAI-compatible APIs only, and document chat and embeddings were removed.
AGPL-3.0 assistant that answers from your files and the web with any model, builds agents and emails scheduled research, with clients for Obsidian, Emacs and WhatsApp.
Know firstIts README does not describe citations or say which search engine it uses.
The facts that decide most choices. Each project's page carries the full panel.
| Project | Best for | Openness | Licence | Runs as | Own machines | Stars |
|---|---|---|---|---|---|---|
| Open WebUI | A private ChatGPT for a whole team | Source available | See LICENSE | pip, Docker, Kubernetes | Yes | 152,426 |
| LibreChat | Every provider in one MIT platform | OSI | MIT | Docker, Docker Compose | Yes | 44,252 |
| AnythingLLM | Chatting with your documents | OSI | MIT | Desktop app, Docker, Bare metal | Yes | 66,155 |
| Jan | A private ChatGPT on one laptop | OSI | See LICENSE | Desktop app, Local OpenAI-compatible server | Yes | 44,516 |
| Chatbox | One app across desktop and phone | OSI | GPL-3.0 | Desktop app, Mobile app | Yes | 41,795 |
| Cherry Studio | Many providers and 300+ assistants in one desktop app | OSI | AGPL-3.0 | Desktop app | Yes | 51,940 |
| LobeHub | Scheduled, collaborating agents | Source available | See LICENSE | Docker, One-click deploy (Vercel, Zeabur, Sealos and others) | Yes | 82,574 |
| TextGen | The most control over local models | OSI | AGPL-3.0 | Desktop app (portable), pip, One-click installer, Docker | Yes | 47,683 |
| Chat UI | HuggingChat's interface on your own server | OSI | Apache-2.0 | npm, Docker | Yes | 10,953 |
| Khoj | Your notes, documents and the web together | OSI | AGPL-3.0 | Self-hosted server, Hosted app, Browser, Obsidian, Emacs, desktop, phone and WhatsApp clients | Yes | 37,399 |
Licence and star figures read from GitHub on September 17, 2026. Stars are shown as a dated fact and were not used to rank this page.
A self-hosted chat platform for Ollama and OpenAI-compatible models with document search over 9 vector databases, web search, image generation, roles and single sign-on, under a BSD-3 licence with an added branding clause.
Documents over 9 vector databases, web search, image generation, voice, and roles with LDAP, OAuth single sign-on and SCIM, on your own server with Ollama or any OpenAI-compatible model.
Know firstSource available: above 50 users you may not remove its branding without permission or a paid licence.
An MIT self-hosted chat platform that puts Anthropic, OpenAI, Google, Bedrock, Azure, Ollama and any OpenAI-compatible model in one interface, with agents, MCP, a code sandbox, web search, artifacts and multi-user sign-in.
Anthropic, OpenAI, Google, Bedrock, Azure and Ollama in one interface, with agents, MCP, a code sandbox for eight languages, web search, artifacts and sign-in.
Know firstA server to run with Docker, with no desktop or mobile app, and every GitHub release tagged pre-release.
An MIT all-in-one AI app from Mintplex Labs for chatting with your documents with citations, running agents with MCP, and using any of 39 model providers including local llama.cpp and Ollama, as a desktop app or a multi-user Docker server.
MIT, with document chat and source citations, agents with MCP, and 39 model providers including local llama.cpp and Ollama, as a desktop app or a Docker server.
Know firstMulti-user support is Docker only, and telemetry is on until you turn it off.
An Apache-2.0 desktop app from Menlo Research that downloads and runs open models from Hugging Face on your own computer, connects to OpenAI, Anthropic and others when you want, serves a local OpenAI-compatible API, and supports MCP.
Apache-2.0 desktop app that runs open models from Hugging Face on your computer, connects to hosted models when you choose, and serves a local OpenAI-compatible API.
Know firstNo document chat, web search or image generation described, and local quality depends on your RAM.
The GPL-3.0 community edition of a desktop and mobile chat client for ChatGPT, Claude, Gemini and local Ollama models that keeps conversations on your device, with a separate pro edition synced from the same code.
GPL-3.0 client for OpenAI, Claude, Gemini and local Ollama models on Windows, macOS, Linux, iOS and Android, with conversations stored on the device.
Know firstNo document chat, web search or MCP described; a separate pro edition exists.
An AGPL-3.0 desktop client for Windows, macOS and Linux that connects OpenAI, Gemini, Anthropic and local Ollama or LM Studio models, with 300+ preset assistants, side-by-side model chats, file support and MCP servers.
AGPL-3.0 client with OpenAI, Gemini, Anthropic, Ollama and LM Studio, 300+ preset assistants, side-by-side model chats, file support and MCP.
Know firstSpeech, OCR and mobile apps are still on its roadmap.
LobeHub, formerly LobeChat, is an agent workspace with an agent builder, agent groups, scheduled runs, editable memory and a library of over 10,000 tools and MCP plugins, under a community licence that needs a paid licence to distribute a modified version.
Agent builder, agent groups, scheduled runs, editable memory and a library of 10,000+ tools and MCP plugins.
Know firstSource available: distributing a modified version needs a paid licence, and the README names only OpenAI in setup.
An AGPL-3.0 desktop app for local models, formerly text-generation-webui, with five engines including llama.cpp and ExLlamaV3, an OpenAI and Anthropic-compatible API, tool calling, file chat, image generation and LoRA training, fully offline.
AGPL-3.0, five engines including llama.cpp and ExLlamaV3, an OpenAI and Anthropic-compatible API, tools, file chat, image generation and LoRA training, fully offline.
Know firstBuilt for technical users; non-GGUF formats need a 10 GB full install.
Hugging Face's Apache-2.0 chat interface, the code behind HuggingChat, that talks to any OpenAI-compatible API, stores chats in MongoDB, calls MCP tools, and can route each message to a model automatically.
Apache-2.0 SvelteKit app that powers HuggingChat, for any OpenAI-compatible endpoint, with MongoDB storage, MCP tools and a per-message model router.
Know firstOpenAI-compatible APIs only, and document chat and embeddings were removed.
An AGPL-3.0 personal AI that answers from the web and your own documents with any local or hosted model, builds custom agents, sends scheduled research by email, and runs self-hosted or in its maintainers' cloud.
AGPL-3.0 assistant that answers from your files and the web with any model, builds agents and emails scheduled research, with clients for Obsidian, Emacs and WhatsApp.
Know firstIts README does not describe citations or say which search engine it uses.
The questions that settle it, in the order they usually come up.
One person: Jan, AnythingLLM, Chatbox, Cherry Studio or TextGen. A team with sign-in and roles: Open WebUI or LibreChat, or AnythingLLM's Docker version.
Jan, TextGen, Open WebUI with its Ollama image, AnythingLLM, Cherry Studio and Chatbox all run local models. Chat UI connects to any local OpenAI-compatible server.
AnythingLLM, with citations, or Open WebUI's document search. Khoj adds notes from Obsidian and Notion.
Choose an OSI licence: LibreChat, AnythingLLM, Jan, Chatbox, Cherry Studio, TextGen or Chat UI. Open WebUI's branding clause and LobeHub's derivative-work clause need a paid licence for that.
ChatGPT, in the vendor's own words, is "your AI assistant for writing, search, image generation, and more", by OpenAI's own description on its Google Play listing. chatgpt.com. The cases below are where that still wins.
Stars, licence, last commit and last release are read from the GitHub API by a script every week and carry the date they were read. Nobody types them.
What each project does is written from its own README and LICENSE, read in full on the date shown, and phrased as the project's claim. The sources are listed at the end.
That a project is faster, better or cheaper than ChatGPT. A price for anyone. A user count. A roadmap. If a fact is not on this page, it was not in the source.
For a team: Open WebUI, the most complete platform, with documents, web search, roles and single sign-on, under a BSD-3 licence plus a branding clause; or LibreChat under MIT with Anthropic, OpenAI, Google and Ollama built in. For one person: Jan for local models, AnythingLLM for documents, Chatbox for desktop and phone.
Yes. Jan runs open models from Hugging Face on your own computer, and its README says everything runs locally when you want it to. TextGen states zero telemetry and no remote requests. Open WebUI ships a Docker image with Ollama bundled, and AnythingLLM, Cherry Studio and Chatbox run local models too.
AnythingLLM, whose main job is chatting with your documents with source citations, using LanceDB by default. Open WebUI also searches documents across 9 vector databases with hybrid search. Khoj answers from your documents and the web, including Notion and Obsidian notes.
Yes. LibreChat lists Anthropic, OpenAI, Google, Bedrock and Azure. Chatbox and Cherry Studio connect OpenAI, Claude and Gemini with your keys. AnythingLLM lists 39 providers. Open WebUI reaches Claude through an OpenAI-compatible gateway such as OpenRouter.
The software is free under each licence. Open WebUI requires a paid licence only if you remove its branding above 50 users, and LobeHub only if you distribute a modified version. Hosted models cost what the provider charges; local models cost nothing but your hardware.
Chatbox has native iOS and Android apps. Open WebUI installs as a web app on a phone. AnythingLLM has a separate MIT mobile app. Khoj works through a phone client and WhatsApp. Cherry Studio lists mobile apps on its roadmap as work in progress.
Running one of these inside your own environment, with your own data and your own security rules, is the kind of work Reveneau does. Read how a forward deployed engagement works.