TODO: fix gemini function calling to keep up with new pipecat version

annie hall weather bot

This repository demonstrates simple llm function calling with Gemini Multimodal Live and parallel Pipecat pipelines. Concept inspired by the "honest subtitles" scene*. Implementation heavily influenced by simple chatbot and this.

Quick Start

1. Copy env.example to .env and add API keys

   cp server/example.env server/.env

2. Setup server

   cd server
   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Start the server

   python server.py

4. Setup client

   cd client/javascript
   npm install

5. Start the client

   npm run dev

6. Navigate to http://localhost:5173 in your browser and click "Talk to annie hall weather bot".

Server Overview

• README
• Pipecat implementation
• Meat of logic in annie-hall.py

Client Overview

• README
• Basic javascript implementation using Pipecat JavaScript SDK
• No framework dependencies

Important Note

The bot server must be running for any of the client implementations to work. Start the server first before trying any of the client apps.

Requirements

• Python 3.10+
• Node.js 16+
• Daily API key (free)
• Gemini API key (free)
• Modern web browser with WebRTC support

* the content can be problematic, but the concept is solid.