What it is

Chatbox (Community Edition) is a cross-platform desktop client for large language models, written in TypeScript and released under the GPL-3.0 license. It runs on Windows, macOS and Linux, and it connects to a range of providers including OpenAI (ChatGPT), Azure OpenAI, Claude, Google Gemini Pro, Ollama and ChatGLM-6B. The project lives in the AI interaction and interfaces ecosystem, where desktop chat front-ends sit between a user and one or more hosted or locally hosted model APIs. The repository carries 41,767 stars and 4,251 forks across four years of history.

The concrete problem it solves is fragmentation. Instead of juggling a separate browser tab, terminal command and API key for every model vendor, a user configures providers once inside a single application and chats from there. It also keeps data on the device through local data storage, so conversation history does not depend on a third-party account. The Community Edition is open-sourced under GPLv3, and code is regularly synced between this repository and the pro repository in both directions.

Key capabilities

• Support for multiple LLM providers, including OpenAI (ChatGPT), Azure OpenAI, Claude, Google Gemini Pro, Ollama and ChatGLM-6B.
• Access to local models through Ollama, such as llama2, Mistral, Mixtral, codellama, vicuna, yi and solar.
• Image generation with Dall-E-3.
• Local data storage, so conversation data remains on the device.
• Streaming replies that render progressively as the model responds.
• Markdown, LaTeX and code highlighting with syntax support for various programming languages.
• Prompt library and message quoting for reuse and context.
• Team collaboration with shared OpenAI API resources.
• Multilingual interface covering English, Simplified and Traditional Chinese, Japanese, Korean, French, German, Russian and Spanish.

Who uses it and how

• Individual desktop users on Windows 10, macOS 11 (Big Sur) or Ubuntu 20.04 and later, installing a native package and configuring a provider in settings.
• Developers who route prompts to local Ollama models rather than paid hosted APIs.
• Teams that share OpenAI API resources through the team collaboration feature.
• Mobile and web users who access the same client through native iOS and Android apps or a browser-based web version.
• Users producing formatted output, such as LaTeX and highlighted code, inside chat threads.

Getting started

Download the installer for the platform from the releases page — Setup.exe for Windows, an Intel or Apple Silicon build for macOS, or an AppImage for Linux — then install, launch and configure an AI provider such as OpenAI or Claude in settings. An APK is published for Android, and a hosted web version is available at chatboxai.app.

When to use it — and when not to

Choose it when the goal is one interface over many providers with data kept locally and no server-side deployment of your own. A self-hoster does not operate a database, object storage or SMTP service for the desktop client, since it is a local application rather than a hosted service; the trade-off is that provider accounts, API keys and any shared team credentials remain the user's responsibility. The main weakness visible in the facts is maintenance load: the repository carries 1,290 open issues, and the Community Edition depends on code synced from a separate pro repository rather than developed independently.