oChat - Async Persona-Driven Chat Interface

A sleek, modern Flask-based chat UI for AI assistants featuring:

• ⚡ Async model replies
• 🎭 Persona + emotional-state engine
• 💾 Conversation saving/loading
• 🗺️ Right-hand minimap of saved chats
• 🌓 Dark/Light mode toggle
• 🔄 Retryable error bubbles on failed replies
• 🧩 Clean template separation (chat.html)

---

📁 Project Structure

oChat/
│
├── chat_async.py
├── api.py
├── characters.py
│
├── templates/
│   └── chat.html
│
├── saved_conversations.json
└── assistant_personalities/

---

🚀 Getting Started

1. Install Dependencies

pip install flask
pip install asyncio

2. Run the Server

python chat_async.py

3. Access the UI at:

http://localhost:4000

---

✨ Features

UI Features

• dark/light mode
• Saved conversation snapshots
• Auto-scroll + typing indicators
• Retry bubble for failed model responses
• Right sidebar with clickable conversation history
• Clean message bubble animation

Backend Features

• Async model inference routing
• Persona system with emotional state engine
• Context summarization
• Pluggable LLM backends (local or remote)
• Conversation reconstruction and history management

---

🎭 Personas

Personas live in:

assistant_personalities/*.json

Each persona includes:

• core_motives
• delivery_style
• signal_strategy
• audience_affinities

Emotional states live in:

assistant_personalities/emotion_states.json

Select the persona via:

export EMOTIONS=Node

---

💾 Saving & Loading Conversations

• Click 💾 Save to snapshot the current chat.
• Right sidebar shows:
  • name
  • timestamp
  • message count

Stored in:

saved_conversations.json

---

🔧 Backend Customization

Create custom personalities for your chatbot by creating a new JSON configuration in assistant_personalities

---

🧪 Error Handling

If the backend times out or fails:

• A red bubble appears: "Failed to get a reply. Click to retry."
• Clicking it resends the same message.

---