Want to quickly prototype your Large Language Model (LLM) application without a ton of code? Unleash your creativity with PromptFlow, a low-code tool that empowers you to craft executable flowcharts. Seamlessly integrate LLMs, prompts, Python functions, and conditional logic to create intricate workflows. With PromptFlow, you can visualize your ideas and bring them to life, all without getting tangled in code or complex logic.

The core of PromptFlow is a visual flowchart editor that lets you design nodes and establish Connections between them. Each node can represent a Prompt, a Python function, or an LLM. The connections between nodes embody conditional logic, dictating the flow of your program.

When run your flowchart, PromptFlow executes each node in the sequence defined by the connections, transferring text data between nodes as required. If a node returns a value, that value is forwarded to the next node in the flow as a string. More information on the inner workings of PromptFlow can be found in our documentation.

You can install PromptFlow via pip with

pip install promptflow-gui

If you want to install PromptFlow from source, clone this repository and install the dependencies with

python -m pip install -r requirements.txt

Encountered a hiccup? Try this instead:

python -m pip install -r requirements-no-nvidia.txt

If you installed from pip, you can launch PromptFlow with

promptflow

Otherwise, run PromptFlow with Python from the command line:

python run.py 

If you're hitting a roadblock, ensure your PYTHONPATH is set correctly:

export PYTHONPATH=$PYTHONPATH:.

Check out our official docs:

To compile the Sphinx documentation, execute:

cd docs
make html

Then, navigate to docs/build/html/index.html in your browser.

Want to contribute to PromptFlow? Get started by building a node.

Stumbled upon a bug? Don't hesitate to create an issue, open a PR, or let us know on Discord.

We're interested in your feedback! If you've used PromptFlow, please fill out this questionnaire.