15.3K ·GitHub
What is this?
ChuanhuChatGPT is a web-based GUI for ChatGPT and many other LLMs, adding extras like agents, file-based Q&A, local model deployment, web search, and GPT fine-tuning on top of a Gradio interface.
ChuanhuChatGPT is a web-based GUI for ChatGPT and many other LLMs, adding extras like agents, file-based Q&A, local model deployment, web search, and GPT fine-tuning on top of a Gradio interface.Is this a desktop app or a web UI I run myself?It's a self-hosted web app built on Gradio that runs locally (or on a server) and opens in your browser; it can also be installed as a PWA for a more native feel on desktop and mobile.
It's a self-hosted web app built on Gradio that runs locally (or on a server) and opens in your browser; it can also be installed as a PWA for a more native feel on desktop and mobile.How do I get it running?Run the project in a terminal, then copy configexample.json to config.json and fill in your API keys; a browser window opens automatically for chatting.
Run the project in a terminal, then copy configexample.json to config.json and fill in your API keys; a browser window opens automatically for chatting.Do I need API keys for these models?Yes — for cloud models (ChatGPT, Claude, Gemini, etc.) you put your API keys in config.json; locally deployed models like ChatGLM or LLaMA don't need a key but require local setup instead.
Yes — for cloud models (ChatGPT, Claude, Gemini, etc.) you put your API keys in config.json; locally deployed models like ChatGLM or LLaMA don't need a key but require local setup instead.Do all supported models work the same way?No — some models like XMChat and Midjourney don't support streaming responses, unlike the main chat models.
No — some models like XMChat and Midjourney don't support streaming responses, unlike the main chat models.Who is this aimed at, and what languages does it support?It's aimed at anyone who wants a nicer, more feature-rich front-end for LLMs; the project itself and its README are available in Chinese, English, Japanese, Russian, and Korean.
It's aimed at anyone who wants a nicer, more feature-rich front-end for LLMs; the project itself and its README are available in Chinese, English, Japanese, Russian, and Korean.Does it run on all my devices?Yes — the UI is adapted for Windows, macOS, Linux, iOS, and Android, with icons and full-screen layouts tuned for each platform.
Yes — the UI is adapted for Windows, macOS, Linux, iOS, and Android, with icons and full-screen layouts tuned for each platform.What can the built-in 'assistant' agent do?The Chuanhu Assistant feature works like AutoGPT, automatically trying to solve your problem end-to-end rather than just answering a single prompt.
The Chuanhu Assistant feature works like AutoGPT, automatically trying to solve your problem end-to-end rather than just answering a single prompt.Can I deploy it on a server instead of my own machine?Yes — you set servername to 0.0.0.0 and a serverport in config.json to expose it on a server, and you can also enable a public share link via Gradio's share option.
Yes — you set servername to 0.0.0.0 and a serverport in config.json to expose it on a server, and you can also enable a public share link via Gradio's share option.Does it support fine-tuning?Yes, it supports fine-tuning GPT-3.5 directly from the UI, letting you create a more personalized model.
Yes, it supports fine-tuning GPT-3.5 directly from the UI, letting you create a more personalized model.Can I search and manage my chat history?Yes — conversation history is saved automatically, isolated per user, renameable, searchable with regex, and can even be auto-named by the LLM itself.
Yes — conversation history is saved automatically, isolated per user, renameable, searchable with regex, and can even be auto-named by the LLM itself.What are similar projects?Langchain-Chatchat, gpt4free, unsloth, ollama, self-llm, LlamaFactory, AstrBot, Awesome-Chinese-LLM
Langchain-Chatchat, gpt4free, unsloth, ollama, self-llm, LlamaFactory, AstrBot, Awesome-Chinese-LLMWhich topics is it in?chatgpt-api, midjourney, chatglm, gemma, minimax, spark, gpt-4, llama, gradio, pwa, qwen, fine-tuning
chatgpt-api, midjourney, chatglm, gemma, minimax, spark, gpt-4, llama, gradio, pwa, qwen, fine-tuning