Open source alternatives to ChatGPT

What ChatGPT does

ChatGPT is OpenAI’s AI assistant for the web, desktop and mobile. It can search the web, read files and images, analyze data, create images, talk by voice, run deep research, and keep memories and projects. Personal plans are Free, Go, Plus and Pro; organizations use Business, Enterprise and Edu. On the free tier, text chats are unlimited subject to abuse safeguards, while uploads, image creation, voice and data analysis have their own limits.

Why people look for an alternative

• Training on personal plans by default. On Free, Plus and Pro personal workspaces, the “Improve the model for everyone” setting is on by default. Turning it off under Settings > Data controls keeps new conversations out of training, but rating an answer with thumbs up or down can still let the whole conversation be used.
• Different rules for business and API use. By default OpenAI does not train on data from ChatGPT Business, Enterprise, Edu or its API, and workspace admins set retention.
• Where conversations are kept. Chats stay on OpenAI’s systems until you delete them, and deleted chats are erased within 30 days unless kept for security or legal reasons. Temporary chats skip history and training but may be kept up to 30 days.
• Data residency only for some customers. Storage in a chosen region (Europe, the UK, the US, Japan and others) is offered to eligible API and new Enterprise and Edu customers, not personal plans.
• Changing models and limits. OpenAI states that the free tier’s default model and limits can change over time.
• No self-hosted version. ChatGPT runs only on OpenAI’s infrastructure. OpenAI’s open-weight gpt-oss models (Apache 2.0) are separate from it.

An open source chat app keeps the interface and chat history under your control, but answers are only as private as the model behind it: with a hosted API, prompts still go to that provider. Open-weight models run locally keep prompts on your hardware, but need a capable GPU or plenty of memory and are generally less capable than the top hosted models.

What to compare in an alternative

1. Models. Which local runtimes (Ollama, llama.cpp, a built-in engine) and hosted providers it supports.
2. Where history lives. Browser, desktop or a server database you back up yourself.
3. Tools. The ones you rely on: web search, document chat, code execution, image generation, voice, memory, agents, MCP.
4. Single user or team. A personal desktop app, or a web app with accounts, SSO and roles.
5. Telemetry and license. What the app reports by default, and whether everything is under an OSI-approved license.

Which one to choose

For a shared, self-hosted ChatGPT-style service, LibreChat is the closest match: a multi-user web app for OpenAI, Anthropic, Google and local models, with agents, MCP, web search and SSO. AnythingLLM focuses on chatting with your documents and running agents; its MIT-licensed Docker server is the open source edition, while the desktop download comes under separate terms. For a personal desktop client, look at Jan, which can also run models locally, and Chatbox. NextChat is a lightweight client you can deploy on Vercel or Docker, Big-AGI can send one prompt to several models with its Beam feature, and Hugging Face’s Chat UI runs HuggingChat. Khoj is a personal assistant built around your own notes and documents. For fully local use, TextGen is a chat app that runs models on your own machine, and PocketPal AI does the same on a phone. The other clients can use local models served by Ollama or llama.cpp.