README En

<div align="right">  <a title="Chinese" href="../README.md">简体中文</a> | English | <a title="Japanese" href="README_ja.md">日本語</a> | <a title="Russian" href="README_ru.md">Russian</a> | <a title="Korean" href="README_ko.md">한국어</a> </div> <h1 align="center">川虎 Chat 🐯 Chuanhu Chat</h1> <div align="center"> <a href="https://github.com/GaiZhenBiao/ChuanhuChatGPT"> </a> <p align="center"> <h3>Lightweight and User-friendly Web-UI for LLMs including ChatGPT/ChatGLM/LLaMA</h3> <p align="center"> <p align="center"><b>New:</b> Now supports GPT-5 family (GPT-5, GPT-5-mini, GPT-5-nano) with 400k context and up to 128k output tokens.</p>

  <a href="https://github.com/GaiZhenbiao/ChuanhuChatGPT/blob/main/LICENSE">
    
  </a>
  <a href="https://gradio.app/">
    
  </a>
  <a href="https://t.me/tkdifferent">
    
  </a>
  <p>
    Compatible with GPT-4 · Chat with files · LLMs local deployment · Web search · Chuanhu Agent ·  Fine-tuning
  </p>
  <a href="https://www.youtube.com/watch?v=MtxS4XZWbJE"><strong>Video Tutorial</strong></a>
    ·
  <a href="https://www.youtube.com/watch?v=77nw7iimYDE"><strong>2.0 Introduction</strong></a>
    ·
  <a href="https://www.youtube.com/watch?v=x-O1jjBqgu4"><strong>3.0 Introduction & Tutorial</strong></a>
||
  <a href="https://huggingface.co/spaces/JohnSmith9982/ChuanhuChatGPT"><strong>Online trial</strong></a>
  	·
  <a href="https://huggingface.co/login?next=%2Fspaces%2FJohnSmith9982%2FChuanhuChatGPT%3Fduplicate%3Dtrue"><strong>One-Click deployment</strong></a>
</p>

</p> </div>

✨ 5.0 Major Update!

<sup>New!</sup> An all-new user interface! So exquisite that it doesn't look like Gradio, it even has a frosted glass effect!

<sup>New!</sup> Adapted for mobile devices (including perforated/bezel-less phones), the hierarchy is clearer.

<sup>New!</sup> The history is moved to the left for easier use. And supports search (with regular expressions), delete, and rename.

<sup>New!</sup> Now you can let the large model automatically name the history (Enabled in the settings or configuration file).

<sup>New!</sup> Chuanhu Chat can now be installed as a PWA application for a more native experience! Supported on Chrome/Edge/Safari etc.

<sup>New!</sup> Icons adapted for all platforms, looking more comfortable.

<sup>New!</sup> Supports Finetune (fine-tuning) GPT 3.5!

Supported Models

API Callable Models	Remarks	Locally Deployed Models	Remarks
ChatGPT(GPT-5, GPT-4, GPT-4o, o1)	supports fine-tune gpt-3.5	ChatGLM (ChatGLM2) (ChatGLM3)
Azure OpenAI		LLaMA	Support Lora models
Google Gemini Pro		StableLM
iFlytek Starfire Cognition Large Model		MOSS
Inspur Yuan 1.0		Qwen
MiniMax
XMChat	Not support streaming
Midjourney	Not support streaming
Claude	✨ Now supports Claude 3 Opus and Sonnet, Haiku will be supported as soon as it is released
DALL·E 3

Usage Tips

💪 Powerful Functions

Chuanhu Assistant: Similar to AutoGPT, automatically solves your problems;
Online Search: Is ChatGPT's data too old? Give LLM the wings of the internet;
Knowledge Base: Let ChatGPT help you speed read quantumly! Answer questions based on files.
Local LLM Deployment: One-click deployment, get your own large language model.
Fine-tuning: Support fine-tuning GPT-3.5, make your own model!
Custom Model: Flexibly customize models, such as docking with local inference services.

🤖 System Prompt

The system prompt can effectively enable role-playing by setting prerequisite conditions;
ChuanhuChat presets Prompt templates, click Load Prompt Template, choose the Prompt template collection first, then choose the Prompt you want in the list below.

💬 Basic Conversation

If the answer is not satisfactory, you can try the Regenerate button again, or directly Delete this round of conversation;
Input box supports line breaks, press <kbd>Shift</kbd> + <kbd>Enter</kbd> to make one;
Using the <kbd>↑</kbd> <kbd>↓</kbd> arrow keys in the input box, you can quickly switch between send records;
Generating a new conversation every time is too cumbersome, try the single-dialogue function;
The small button next to the answer bubble not only allows one-click copy, but also lets you view the original Markdown text;
Specify the answer language, so that ChatGPT will always reply in a certain language.

📜 Chat History

Dialogue history will be automatically saved, you won't have to worry about not being able to find it after asking;
Multi-user history isolation, only you can see them;
Rename chat, easy to find in the future;
<sup>New!</sup> Magically auto-name the chat, let LLM understand the conversation content, and automatically name the chat for you!
<sup>New!</sup> Search chat, supports regular expressions!

🖼️ Small and Beautiful Experience

Self-developed Small-and-Beautiful theme, gives you a small and beautiful experience;
Automatic light and dark color switching, gives you a comfortable experience from morning till night;
Perfectly rendering LaTeX / tables / code blocks, supports code highlighting;
<sup>New!</sup> Non-linear animations, frosted glass effect, so exquisite it doesn't look like Gradio!
<sup>New!</sup> Adapted for Windows / macOS / Linux / iOS / Android, from icon to screen adaptation, gives you the most suitable experience!
<sup>New!</sup> Supports PWA app installation for an even more native experience!

👨‍💻 Geek Functions

<sup>New!</sup> Supports Fine-tuning gpt-3.5!
Plenty of available LLM parameters to adjust;
Supports API-host switching;
Supports custom proxies;
Supports multiple api-key load balancing.

Deployment to the server: Set in config.json "server_name": "0.0.0.0", "server_port": <your port number>,.
Obtain public link: Set in config.json "share": true,. Note that the program must be running to access it through public links.
Use on Hugging Face: It's recommended to Duplicate the Space in the top right corner before using, the App response might be faster.

Quick Start

Execute the following commands in the terminal:

shell

git clone https://github.com/GaiZhenbiao/ChuanhuChatGPT.git
cd ChuanhuChatGPT
pip install -r requirements.txt

Then make a copy of config_example.json, rename it to config.json, and then fill in your API-Key and other settings in the file.

shell

python ChuanhuChatbot.py

A browser window will automatically open, at this point you can use Chuanhu Chat to chat with ChatGPT or other models.

Note

Please check our wiki page for detailed instructions.).

Troubleshooting

When you encounter problems, you should try to manually pull the latest changes<sup>1</sup> and update dependencies<sup>2</sup> first, then retry. Steps are:

Click on the Download ZIP button on the website, download the latest code and unzip to replace, or
shell
```
git pull https://github.com/GaiZhenbiao/ChuanhuChatGPT.git main -f
```
Try to install dependencies again (the project might have new dependencies)
```
pip install -r requirements.txt
```

Generally, you can solve most problems by following these steps.

If the problem still exists, please refer to this page: Frequently Asked Questions (FAQ)

This page lists almost all the possible problems and solutions. Please read it carefully.

More Information

More information could be found in our wiki:

Starchart

Contributors

🐯 If you find this project helpful, feel free to buy me a coke or a cup of coffee~