CLI Chatbot with Memory using Replicate API

This is a command-line chatbot leveraging Replicate.com to run the openai/gpt-4o-mini language model with persistent conversation memory and speech synthesis.

---

Features

• Persistent conversation saved in JSON files
• Start, end, and control chat sessions with commands
• Auto and manual text-to-speech (TTS) playback
• Speech audio saved to ./services/sounds
• Commands to stop speech and clear audio files
• Easily configurable via .env file

---

Setup Instructions (Windows PowerShell)

1. Clone or download this repository

git clone https://github.com/GitHubNexon/FINCH.git

2. Create .env file

Create a .env file in the project root directory with the following content:

REPLICATE_API_TOKEN=your_replicate_api_token_here
MODEL_NAME=openai/gpt-4o-mini

You can get your Replicate API token here: https://replicate.com/account/api-tokens

3. Create Virtual Environment

python -m venv venv

4. Activate Virtual Environment

.�env\Scripts\Activate

5. Install Required Packages

pip install -r requirements.txt

6. Run the Chatbot

python chat.py

---

Chat Commands

Command	Description
/start	Start a new conversation session
/end	End the current conversation and clear memory
/stop	Stop speech playback
/clear	Clear saved audio files from sounds folder

---

Notes

• Conversations are stored in JSON files to allow persistent memory between sessions.
• Speech audio files are stored in ./services/sounds. Use /clear command to remove them.
• Ensure your .env file is properly configured before running the chatbot.

---