Talki

Talki: it means Talk in English and Bunny in Korean! Talki is a voice-driven chat application with frontend and backend components. The frontend is a Next.js app that provides a conversational UI and media upload endpoints; the backend is a Flask-based service that handles audio processing, text-to-speech and speech-to-text, and image generation.

Demo video

https://youtu.be/gc4kM7IjK_0

This repository contains two main folders:

• Talki-back/: Python backend (Flask) — audio generation, speech-to-text (STT), and image generation utilities.
• Talki-front/: Next.js frontend — user interface, API routes that call the backend, and chat components.

Table of contents

• Installation
• Backend (Talki-back)
• Frontend (Talki-front)
• Project structure
• Troubleshooting
• Contributing

Installation

Prerequisites

• Node.js (16+) and npm or Yarn for the frontend
• Python 3.10+ for the backend
• Optional: Google Cloud credentials for TTS/STT if used (Talki-back/my-gcp-credentials.json)

Clone the repository and change into the workspace root:

git clone <your-repo-url>
cd Talki

Backend (Talki-back)

1. Create and activate a Python virtual environment, then install dependencies:

cd Talki-back
python -m venv .venv
.venv\Scripts\activate    # Windows (cmd.exe)
# or: source .venv/bin/activate    # macOS / Linux
pip install -r requirements.txt

2. Credentials

• If using Google Cloud TTS/STT, place your service account JSON in Talki-back/my-gcp-credentials.json and set the environment variable:

set GOOGLE_APPLICATION_CREDENTIALS=%CD%\\my-gcp-credentials.json

3. Run the backend

python app.py

The backend serves endpoints used by the frontend and writes generated media into Talki-back/static/.

Frontend (Talki-front)

1. Install dependencies and run the Next.js app:

cd ..\\Talki-front
npm install
npm run dev

2. By default the frontend expects the backend to be reachable via the API routes defined in app/api/* (the frontend includes server-side API routes that may proxy or call the backend). If the backend runs on a different host/port, update the frontend API base URL (look for fetch calls or environment variables in Talki-front/app/api routes).

Project structure

• Talki-back/

  • app.py — Flask application entrypoint
  • services.py — backend helper functions used by the app
  • requirements.txt — Python dependencies
  • my-gcp-credentials.json — (optional) Google Cloud credentials (NOT checked into VCS)
  • static/ — generated audio and image assets

• Talki-front/

  • app/ — Next.js app
  • app/api/ — Next.js API routes
  • app/components/ — React components used by the UI

Troubleshooting

• If you see errors related to Google Cloud authentication, confirm GOOGLE_APPLICATION_CREDENTIALS points to a valid service account JSON and that required APIs are enabled.
• For CORS or proxy issues, ensure the frontend can reach the backend host and port, or configure a proxy in development.

Contributing

Contributions welcome. Create issues or pull requests for bug fixes or feature additions. Keep secrets and credentials out of commits.