Copyright (c) 2024-2025 led-mirage
This is an app that allows you to chat with AI. A unique feature of this app is that it reads questions and answers aloud.
The available AIs are as follows:
- OpenAI GPT (Azure OpenAI Service is also available)
- Google Gemini
- Anthropic Claude
ZundaGPT2_1.6.0en.mp4
Starting from version 1.6.0, ZundaGPT2 supports both English and Japanese, allowing for flexible use depending on your language preference.
To change the language, modify the "language" setting in the appConfig.json file:
"language": "en"You can switch between "en" for English and "ja" for Japanese.
- Windows 11 Pro 23H2、24H2
- Python 3.12.0
- VOICEVOX 0.22.3
- A.I.VOICE Editor 1.4.10.0
- COEIROINK v.2.3.4
While this app itself is free, one of the following API keys is required to operate it.
To use the OpenAI API, you will need an account with OpenAI and must register for API access (including billing and API key creation).
As of version 0.7.0, this app also supports the Google Gemini API.
As of May 19, 2024, the Google Gemini API offers a free plan, making it more accessible compared to the OpenAI API.
With version 1.4.0, support has been added for the Anthropic API (Claude series).
To use this API, an account on Anthropic Console and registration for API access (including billing and API key creation) are required.
As of December 29, 2024, the latest model is Claude 3.5 Sonnet.
This app supports the following text-to-speech software. You need to have at least one of them installed on your PC.
- VOICEVOX (Free)
- COEIROINK (Free)
- A.I.VOICE (Paid)
- Google Text-to-Speech (Free) ... Requires FFmpeg
- SAPI5 (Free) ... No installation required
Once you're ready, move on to the next step for execution.
You need to register the API key for OpenAI, Google Gemini API, or Anthropic API as an environment variable in your OS.
| AI | Variable Name | Value |
|---|---|---|
| OpenAI | OPENAI_API_KEY | The API key obtained from OpenAI |
| Google Gemini | GEMINI_API_KEY | The API key obtained from Google |
| Anthropic Claude | ANTHROPIC_API_KEY | The API key obtained from Anthropic |
For Windows, search "Edit environment variables" in the search bar to open the settings window, where you can add user environment variables.
Create a folder for the project at a location of your choice.
Download ZundaGPT2.ZIP from the following link and extract it into the folder you created.
https://github.com/led-mirage/ZundaGPT2/releases/
Double-click ZundaGPT2.exe to launch the app.
If you prefer not to display the splash screen at startup, use ZundaGPT2.ns.exe instead.
This executable file (EXE) is created using a library called PyInstaller. Unfortunately, it is often mistakenly flagged as malware or a virus by antivirus software. This issue is widely discussed online, indicating it is a common occurrence, but a fundamental solution has not yet been found.
Rest assured, there are no malicious programs included in this app. However, if you have concerns, please execute it using the "Running with Python" method described below.
First, set up an environment where Python 3.12.0 is operational. It might work on other versions, but they haven't been tested yet.
I use pyenv-win + venv to create a virtual environment for development, so that method works as well.
Create a folder for the project at a location of your choice.
Open a terminal or command prompt, and navigate to the project folder you created.
Download the ZIP file and extract it into the folder.
Alternatively, if you can use Git, run the following command to clone the repository:
git clone https://github.com/led-mirage/ZundaGPT2.gitRun the following command from the command prompt to install the necessary libraries:
pip install -r requirements.txtTo launch the app, execute the following command from the command prompt:
python app\main.pyIt can be handy to create a startup batch file like the one below:
start pythonw app\main.pyIf you're using a Python virtual environment, you can set it up like the example below:
call venv\scripts\activate
start pythonw app\main.pyYou can select different AI personalities by clicking the ⚙️ button in the top-right corner of the screen.
While there are several default character settings available, you can create your own custom characters by copying and editing the existing setting files.
Character setting files (settings_xxx.json) are stored in the settings folder. Simply copy an existing file and modify it to create your own character.
For detailed information about character configuration, please refer to this guide.
This app is free, but using the OpenAI API incurs additional charges (there is a free trial tier available). So be careful not to overuse it. I recommend regularly checking your current usage on the OpenAI website to manage it yourself.
Moreover, if you don't enable auto-recharge settings, you should only be charged the amount you have precharged, so there shouldn't be too much to worry about. However, on the OpenAI website, you can set a monthly usage limit, so it's a good idea to use those features to avoid unexpected expenses.
As mentioned in this document, there is currently a free tier available for the Google Gemini API. Therefore, for the most part, you can use the app within this free tier. However, if you want to use it more extensively, consider upgrading to a paid plan. Just like with OpenAI, be cautious of overuse if you switch to a paid plan.
Using the Anthropic API also incurs separate charges (pay-as-you-go). As of December 29, 2024, there seems to be no free tier available. You can use the API by charging any amount you like on your credit card. However, as with other APIs, be cautious of excessive usage.
The API keys for OpenAI, Google Gemini, and Anthropic are for your use only, so don't share them with others. If they leak, they might be used without your permission by malicious parties. If your key is compromised, simply delete the current API key via the OpenAI, Google, or Anthropic portal and create a new one.
However, with OpenAI, if you only have one API key, it seems you can't delete it until you create a new one. This is apparently a system limitation with OpenAI, and personally, I wish it was better designed. I hope for improvements in the future, but as a last resort, removing your payment information (credit card details) might be an option.
Above all, please pay close attention to your API keys and usage fees.
So far, no particular issues have been found, but bugs are inherently potential issues and can never be entirely eliminated. Additionally, please understand that the copyright holder cannot be held liable for any damages you may incur as a result of these bugs. Use the app with this understanding.
Homepage： https://github.com/r0x0r/pywebview
License：BSD-3-Clause license
Homepage： https://github.com/openai/openai-python
License：Apache License 2.0
Homepage： https://github.com/google-gemini/generative-ai-python
License：Apache License 2.0
Homepage： https://github.com/anthropics/anthropic-sdk-python
License：MIT license
Homepage： https://github.com/pndurette/gTTS
License：MIT License
Homepage： https://requests.readthedocs.io/en/latest/
License：Apache License 2.0
Homepage： https://people.csail.mit.edu/hubert/pyaudio/
License：MIT License
Homepage：https://github.com/pythonnet/pythonnet
License：MIT License
Homepage：https://github.com/Mimino666/langdetect
License：Apache License 2.0
Homepage：https://github.com/jiaaro/pydub
License：MIT License
Homepage：https://github.com/mhammond/pywin32
License：Python Software Foundation License (PSF)
Homepage： https://github.com/mathjax/MathJax
License：Apache License 2.0
Homepage：https://github.com/highlightjs/highlight.js
License：BSD-3-Clause license
Homepage：https://github.com/markedjs/marked
License：MIT license
Homepage：https://github.com/julkue/mark.js
License：MIT license
Homepage：https://fontawesome.com/ License：Icons: CC BY 4.0, Fonts: SIL OFL 1.1, Code: MIT License
© 2024-2025 led-mirage
This application is released under the MIT License. For more details, please refer to the LICENSE file included in the project.