NextChat
Open-source chat app for talking to many AI models, with a web and desktop version.
- Stars
- 88.9k
- Forks
- 58.6k
- Licence
- MIT
- Language
- TypeScript
- Last update
- 17 days ago
- Type
- App
What it is
NextChat is a chat interface for AI language models such as GPT, Claude, Gemini and DeepSeek. You can use it in a browser as a web app, or install it as a desktop app on Windows, macOS and Linux. The README also describes self-hosting it, for example on Vercel, and connecting it to self-deployed models.
- Chat with a hosted model such as GPT or Gemini from a browser tab
- Deploy your own copy of the chat interface with a one-click Vercel button
- Connect the app to a self-hosted model via RWKV-Runner or LocalAI
- Create and reuse prompt templates called masks
- Share conversations as images
Great for
- People who want one chat window for several AI models
- Developers who want a self-hosted chat front end for their own API keys
- Users who want local storage of chat data in the browser
- Teams interested in the README's enterprise edition for private deployment
Not ideal for
- People who want a fully offline assistant, since the README centers on connecting to hosted or self-deployed model services
- Users who need the iOS app's source code, since the README says it is coming soon
- Anyone expecting a guaranteed no-cost service, since the README mentions pay-as-you-go and a sponsor platform
How to get it
- 1
Use it in the browser (or self-host)
Clone the repository
git clone https://github.com/ChatGPTNextWeb/NextChat - 2
Run it
- Open the web app at https://app.nextchat.club or deploy your own copy with the Vercel button in the README.
- Set the OPENAI_API_KEY environment variable if you deploy your own copy, as the README's deploy link asks for it.
- Set a page password through the CODE variable so only you can use your deployment.
- Open the app and start chatting with a model.
- 3
Get the most out of it
- Use prompt templates (masks) to save reusable chat setups
- Set ENABLE_MCP=true before building to enable MCP support, as the README describes
- Connect a self-hosted model through RWKV-Runner or LocalAI if you do not want to use a hosted provider
- Enable Markdown features such as LaTeX, Mermaid diagrams and code highlighting for technical chats
Downloads
Strengths
- Runs on web, Windows, macOS, Linux and iOS according to the README
- Stores data locally in the browser, per the README's privacy note
- Open source under the MIT license with a large community
Trade-offs
- The README is truncated and says little about setup beyond Vercel deployment and build flags
- Several features, such as MCP support, require building with environment flags
- Some links and release downloads point to a different repository name than the one listed
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| NextChat (this) | 88.9k | App | beginner | 17 days ago | People who want one chat window for several AI models |
| LibreChat | 46.6k | App | intermediate | 4 days ago | Teams that want one shared chat interface across several AI providers |
| openwhispr | 10.5k | App | beginner | 4 days ago | People who want dictation in any app without sending audio to a cloud service |
| NotionNext | 11.9k | App | intermediate | 4 days ago | Writers who already keep their posts in Notion and want a website without moving content |
| DevToys | 32.2k | App | beginner | 1 months ago | Developers who often decode, encode, or format small pieces of data |
| taxonomy | 19.3k | App | advanced | 7 months ago | Developers learning the Next.js App Router and server components |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.