ArgueBot

A Gradio web app that lets you upload text to train a chatbot's speaking style, chat with it via your local LLM (Ollama), and hear replies spoken through your local Chatterbox TTS Docker server.

Features

• Style training — upload .txt, .md, .csv, .pdf or paste text; more text = better style matching (RAG)
• Chat — bot avatar, conversation history, auto-spoken replies
• Chatterbox TTS — uses your Docker chatterbox-tts-server on port 8004
  • Predefined voices (Abigail, Olivia, etc.)
  • Clone voices from your reference audio library (AlexJones, kryten2, etc.)
• Settings — LLM URL/model, Chatterbox URL, voice mode, TTS parameters, persona, avatar
• Remote access — python app.py --share creates a temporary gradio.live URL (same as podcaster2)

Prerequisites

Service	This machine
Ollama	http://127.0.0.1:11434 (default LLM)
Chatterbox TTS	Docker container chatterbox-tts-server-cu128 on http://127.0.0.1:8004
Python	3.11+

Port 8080 on this machine is used by Frigate, not the LLM — ArgueBot defaults to Ollama on 11434.

Quick Start

cd ArgueBot

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

cp .env.example .env   # optional — edit defaults

# Local only
python app.py

# Temporary public URL (gradio.live)
python app.py --share

Open http://localhost:7860

Usage

1. Training tab

Upload writing samples (social posts, transcripts, articles). The bot retrieves relevant excerpts when answering to match your style.

2. Chat tab

Type a question. ArgueBot crafts a styled reply via Ollama and speaks it through Chatterbox.

3. Settings tab

• LLM: http://127.0.0.1:11434, model e.g. qwen2.5:7b
• Chatterbox: http://127.0.0.1:8004
• Voice mode:
  • clone — use a reference audio file from your Chatterbox library (e.g. kryten2.mp3)
  • predefined — built-in voice (e.g. Olivia.wav)
• Click Refresh voices to pull the latest list from Chatterbox

Voices and reference audio are managed in the Chatterbox UI at http://127.0.0.1:8004 — ArgueBot does not duplicate that.

Chatterbox Docker

ArgueBot expects Chatterbox-TTS-Server running locally:

# Already running on this machine as:
docker ps | grep chatterbox
# chatterbox-tts-server-cu128   0.0.0.0:8004->8004/tcp

Manage voices and models via the Chatterbox UI at http://127.0.0.1:8004

CLI options

python app.py                    # local
python app.py --share            # public gradio.live link
python app.py --server-port 7861
python app.py --server-name 0.0.0.0

Project structure

ArgueBot/
├── app.py              # Gradio UI
├── prompts.py          # Style-aware system prompts
├── services/
│   ├── chatterbox.py   # Chatterbox-TTS-Server client
│   ├── corpus.py       # Text ingestion + chunking
│   ├── llm.py          # Ollama / OpenAI-compat client
│   ├── rag.py          # TF-IDF style retrieval
│   └── settings_store.py
└── data/               # corpus, avatar, SQLite (gitignored)

License

MIT