Wordflow

Social and customizable AI writing assistant tool ✍️

🚀 Wordflow Demo	📺 Demo Video	📖 Research Paper

What is Wordflow?

Wordflow is a social and customizable AI writing assistant tool! With Wordflow, you can easily write and run AI prompts with different large language models, including both external models (e.g., GPT 5.4 and Gemini Pro) and local models (e.g., Llama 2 and Phi 2). You can also discover and share your favorite prompts with the community.

✅	Store and run your favorite AI prompts
✅	Support different external LLMs (e.g., GPT 5.4 and Gemini Pro)
✅	Support multiple on-device LLMs (e.g., Llama 2 and Phi 2)
✅	Powerful customization (e.g., temperature, templates)
✅	Discover and share community prompts

Features

Demo Video

Click to see the demo video!

Get Started

Standalone Version

To use Wordflow, visit: https://poloclub.github.io/wordflow/.

If you use macOS, we highly recommend you creating a Web App using Safari. Open Wordflow in Safari, and then from the menu bar, choose File > Add to Dock.

Note that currently Safari doesn't support WebGPU. If you want to use local LLMs, you can use Chrome to save Wordflow as a Chrome app. In Chrome, at the top right, click More > More Tools > Create shortcut.

Google Doc Add-on

If you use Google Doc, we also provide a Wordflow Add-on that you can install from the Google Workspace Marketplace. The source code of the add-on is in this repository.

Developing Wordflow

Clone or download this repository:

git clone git@github.com:poloclub/wordflow.git

Install the dependencies:

npm install

Then run Wordflow:

npm run dev

Navigate to localhost:3000. You should see Wordflow running in your browser :)

Credits

Wordflow is created by Jay Wang, Aishwarya Chakravarthy, David Munechika, and Polo Chau.

Citation

To learn more about Wordflow and social prompt engineering, check out our research paper.

@article{wangWordflowSocialPrompt2024,
  title = {Wordflow: {{Social Prompt Engineering}} for {{Large Language Models}}},
  shorttitle = {Wordflow},
  author = {Wang, Zijie J. and Chakravarthy, Aishwarya and Munechika, David and Chau, Duen Horng},
  year = {2024},
  url = {http://arxiv.org/abs/2401.14447},
  urldate = {2024-01-29},
  archiveprefix = {arxiv},
  journal = {arXiv 2401.14447}
}

License

The software is available under the MIT License.

Contact

If you have any questions, feel free to open an issue or contact Jay Wang.