ChatALL (Chinese name: Qidao) is a client for concurrent chatting with AI bots based on large language models. Its core purpose is to let users send the same prompt to multiple AI bots at the same time and compare their answers, so they can quickly find the best result.
Main capabilities:
- Concurrently send prompts to multiple AI bots, including ChatGPT, Bing Chat, Bard, Alpaca, Vicuna, Claude, ChatGLM, MOSS, iFlytek Spark, ERNIE Bot, and more.
- Fast prompt mode: send the next prompt without waiting for the previous request to finish.
- Save chat history locally to protect privacy.
- Highlight liked answers and delete bad answers.
- Enable or disable any bot at any time.
- Switch between single-column, two-column, or three-column views.
- Automatically update to the latest version.
- Dark mode (contributed by @tanchekwei).
- Keyboard shortcuts; press Ctrl + / to view all shortcuts (contributed by @tanchekwei).
- Multiple chats (contributed by @tanchekwei).
- Proxy settings (contributed by @msaong).
- Prompt management (contributed by @tanchekwei).
- Supports multiple languages: Chinese, English, German, French, Russian, Vietnamese, Korean, Japanese, Spanish, and Italian.
- Supports Windows, macOS, and Linux.
How to use:
- Fromhttps://github.com/ai-shifu/ChatALL/releases下载安装。Windows下载*-win.exe文件并安装；Apple For Silicon Mac (M1, M2 CPU), download the *-mac-arm64.dmg file; for other Macs, download the *-mac-x64.dmg file; Homebrew users can also install via Homebrew. For Debian-based distributions, download the .deb file and double-click to install; for Arch-based distributions, it can be cloned and installed from the AUR; for other distributions, download the .AppImage file, make it executable, and run it.
- ChatALL is a client rather than a proxy, so users must have available accounts and/or API tokens for the corresponding bots, as well as a reliable network connection.
- If you encounter issues, try refreshing (Ctrl + R or ⌘ + R), restarting, logging in again, creating a new chat, or resetting ChatALL (delete the configuration directory for the corresponding system: on Windows, C:\Users<user>\AppData\Roaming\chatall; on Linux, /home/<user>/.config/chatall/; on macOS, /Users/<user>/Library/Application Support/chatall/).
Use cases:
- LLM experts: those who want to find the best answer or creation from LLMs.
- LLM researchers: those who want to intuitively compare the strengths and weaknesses of different LLMs in different domains.
- LLM application developers: those who want to quickly debug prompts and find the base model with the best performance.
Limitations and notes:
- Web-access AI bots (marked as "Web Access") have lower reliability and often face stability issues because service providers regularly update web interfaces and security measures. Such connections rely on reverse engineering, are difficult to maintain, and are prone to unexpected interruptions. It is recommended to use bots that provide API access whenever possible.
- All chat history, settings, and login data are stored on the local computer.
- ChatALL collects anonymous usage data to help improve the product, including: which AI bots were prompted and the prompt length (excluding prompt content); response length and which response was deleted/highlighted (excluding response content).
- Planned features: deploy the frontend to GitHub Pages.
