🤖 AI Chatbot Backend (Python + Ollama + TTS)

This project uses a fully open-source AI stack to provide:

💬 Streaming AI chat replies

🎤 Voice output (Text-to-Speech)

🧠 Local LLM via Ollama

🌐 FastAPI backend

📦 Requirements

Ollama (Local LLM Server)

Install Ollama:

• 👉 https://ollama.com

After installing, pull a model:

• ollama pull llama3

Start Ollama server:

• ollama serve

• (Default runs on http://localhost:11434)

Python Dependencies

Python (Recommended) Python 3.11.x -> (Required for TTS compatibility)

Install required packages:

• pip install fastapi uvicorn requests TTS

🚀 Running the Backend

Start FastAPI server:

• python -m uvicorn main:app --reload

Backend will run on:

• http://127.0.0.1:8000

🔁 Chat Streaming Endpoint

• POST /chat-stream

• Streams AI response from Ollama in real-time.

🔊 Voice (Text-to-Speech)

• This project uses Coqui TTS (open-source) for natural AI voice.

Example model:

• tts_models/en/ljspeech/glow-tts

Voice is generated after full AI message is received.

• 🌍 Web Support (CORS Enabled)

• FastAPI is configured with CORS to support.

🧠 Stack Overview

• FastAPI -> Backend server
• Ollama -> Local LLM (AI brain)
• Coqui TTS -> Voice generation
• Flutter -> Frontend UI (https://github.com/darttechwala/ChatBot)

✅ Features

• Fully offline capable (local AI)

• Streaming chat responses

• Voice replies

• Multi-platform (Android, iOS, Web, macOS, Windows)

• Open-source stack

⚠ Notes

Ollama must be running before starting backend