Gemini Chat Interface

Warning

Work-in-progress: This project is currently in a very early stage. Only the minimum core functionality is implemented, and many features are still missing or incomplete.

A simple, open-source Next.js web interface for interacting with Google's Gemini AI API. It provides a ChatGPT-like chat experience with basic streaming response functionality, using PostgreSQL to store chat history.

Features

• Interactive, ChatGPT-inspired chat interface
• Integration with Google's Gemini AI API
• Basic streaming responses
• Persistent conversation history (PostgreSQL)
• Support for free and paid Google API keys
• Markdown and code syntax highlighting
• Modern UI built with Next.js and Tailwind CSS

Installation

Prerequisites

• Docker
• Google Gemini API key (get here)

1. Clone the repository:

git clone https://github.com/W4D-cmd/QuantaGem.git

2. Environment Configuration

Create a new file .env.local in the root of the repository. Copy and paste the content of the .env file into it and set your API keys.

Note

If you do not have multiple Google accounts or wish to only use the free API simply put the same key for both entries.

FREE_GOOGLE_API_KEY="your_free_google_api_key"
PAID_GOOGLE_API_KEY="your_paid_google_api_key"

You must also set JWT_SECRET to a random, cryptographically strong string. This secret is vital for securing user sessions and should be at least 32 characters (256 bits) long. You can generate a suitable value using node -e "console.log(require('crypto').randomBytes(32).toString('base64'))" and add it to your .env.local file.

JWT_SECRET="your_jwt_secret"

3. Customizing the Speech-to-Text (STT) Model

The application uses the medium as the default model for Speech-to-Text transcription. You can change this to any other model from the Faster Whisper family to balance performance and accuracy according to your hardware and needs.

To change the model, you need to edit the model identifier string in the STT service's source code.

1. Open the file stt-service/main.py.

2. Locate the model_size variable at the top of the file.

   model_size = "medium"
   compute_type = "int8"

3. Replace the string value of model_size (e.g., "medium") with the name of your desired model from the list below (e.g., "distil-large-v3").

4. Save the file and rebuild the Docker container using docker compose up --build for the changes to take effect.

Available Faster Whisper Models

Here is a list of available models, grouped by type. Larger models are more accurate but slower and require more resources.

Standard Models (Multilingual)

• tiny
• base
• small
• medium
• large-v1
• large-v2
• large-v3

English-Only Models (.en)

• tiny.en
• base.en
• small.en
• medium.en

Distilled Models (distil)

• distil-small.en
• distil-medium.en
• distil-large-v2
• distil-large-v3

Running the Application

Inside the cloned repository execute the following command to start up the docker environment including the database and the Next.js app:

docker compose up --build

Open your browser at http://localhost:3000.

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create your feature branch (git checkout -b feature/my-feature).
3. Commit your changes (git commit -am 'Add new feature').
4. Push to the branch (git push origin feature/my-feature).
5. Create a new pull request.

License

Licensed under the MIT License. See LICENSE for details.