NextChat (88,588 Stars) 2026: Lightweight ChatGPT Web UI That Runs on Web, iOS, MacOS, Android and Linux

NextChat (88,588 stars) is the light and fast AI assistant UI supporting Web, iOS, MacOS, Android, Linux and Windows. Here is how to deploy it in one command.

💡 What You Will Learn

NextChat (88,588 stars) is the light and fast AI assistant UI supporting Web, iOS, MacOS, Android, Linux and Windows. Here is how to deploy it in one command.

📜 Table of Contents

• The short answer
• Highlights
• Deploy in one command (Docker)
• Configuring local models
• Practical tips
• FAQ

The short answer

ChatGPTNextWeb/NextChat (88,588 stars, TypeScript) is a light and fast AI assistant UI. It started as a ChatGPT web wrapper and grew into a cross-platform app covering Web, iOS, MacOS, Android, Linux, and Windows - all from one codebase. You bring your own API key and it works.

Highlights

• One-click deploy: Vercel, Docker, or a static export
• Multi-provider: OpenAI, Claude, Gemini, DeepSeek, and any OpenAI-compatible endpoint
• Cross-platform: desktop apps for all major OSes via Tauri
• Prompt library: built-in prompt management and sharing
• Local-first: conversations can be stored locally or in the cloud of your choice

Deploy in one command (Docker)

docker run -d -p 3000:3000 \n  -e OPENAI_API_KEY=sk-... \n  --name nextchat \n  yidadaa/chatgpt-next-web

Open http://localhost:3000. You can also deploy free on Vercel by importing the repo and setting the API key in environment variables.

Configuring local models

Set OPENAI_API_KEY to any key and add a custom endpoint:

-e BASE_URL=http://localhost:11434/v1 \n-e OPENAI_API_KEY=ollama

Then pick a model like qwen2.5:7b in the settings and start chatting with your local LLM.

Practical tips

• Set CODE (access password) if you expose it publicly.
• Use the built-in prompt store to keep frequently used prompts organized.
• The desktop app (download from the releases page) gives you a native window on Windows/macOS/Linux.

FAQ

Is it free? Yes - MIT licensed; you pay only for model API usage.

Does it need a database? No - it is stateless; conversations are stored in browser storage or via optional services.

Can it use Azure OpenAI? Yes - there are settings for Azure endpoints and deployment names.

❓ FAQ

Is it free?

Yes - MIT licensed; you pay only for model API usage.

Does it need a database?

No - it is stateless; conversations are stored in browser storage or via optional services.

Can it use Azure OpenAI?

Yes - there are settings for Azure endpoints and deployment names.