Temporal Streaming with Redis PubSub

A solution for real-time streaming output from Temporal workflows using Redis PubSub. This addresses the limitation that Temporal doesn't natively support streaming responses from an LLM.

Overview

This project demonstrates how to implement real-time streaming from Temporal workflows by using Redis PubSub as a communication channel. This demonstration focuses specifically on streaming LLM output back to the user via CLI, while also persisting results to the Temporal Activity.

How It Works

1. A Temporal workflow is defined that executes an activity
2. The activity streams its output to a Redis PubSub channel character by character
3. A client subscribes to the channel and displays the stream in real-time in the terminal
4. The workflow completes once all streaming is finished

Requirements

• anthropic
• python-dotenv
• UV package manager (for dependency management)
• Python 3.10+
• Redis server
• Temporal server

Installation

# Create a virtual environment using UV (recommended)
uv venv

# Activate the virtual environment
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Install dependencies with UV
uv sync

Usage

1. Start a Redis server in a separate terminal:

redis-server # Validate it's running in another terminal by running 'redi-cli ping' --> response should be 'pong'

2. Start a Temporal server in a separate terminal:

temporal server start-dev # Open the Temporal UI on localhost:8233

3. Start the Worker from the original virtual environment:

python worker.py

4. Start the Workflow from a new terminal with the virtual environment

python starter.py

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT