AI static site generator. Given a folder of pseudocode, JennGen generates a folder of real code. It works best with HTML, CSS, and JavaScript, but it can translate any language or text-based file format. For example, you can use a Markdown bullet list of instructions to generate a Dockerfile. Given good examples, you should be able to generate Terraform, Python, or JSON as needed for functionality beyond a basic static site.
Get your OpenAI key from https://platform.openai.com/api-keys.
Build it once:
export OPENAI_API_KEY=<YOUR_API_KEY>
npx jenngen .
Build it continuously:
npx jenngen . --watch --server
open localhost:3000
You can use a local model with Ollama. This is useful if you want to use an open source model that is not available on OpenAI's API. You can also use it to avoid the API rate limits. See the Ollama model library for the list of available models.
brew install ollama
brew services start ollama
ollama pull mistral
export JENNGEN_OLLAMA_MODEL=mistral
npx jenngen . --server
http://localhost:3000
JennGen uses OpenAI's GPT4 to translate pseudocode into real code using custom prompting and examples. You can add a .jenngen file to your project to provide further instructions.
<your project>
├── .jenngen_cache
│ └── <cached files>
├── .dist
│ ├── index.html
│ ├── style.css
│ └── script.js
├── .jenngen
├── index.html
├── style.css
└── script.js
The .jenngen file is a plaintext file that contains instructions for JennGen. It can be used to provide examples, add custom prompts, and more.
npx jenngen <source folder> [options]--docs- Experimental: use RAG to generate docs from markdown files--force- always generate files, even if they haven't changed--server- start a server to serve the output folder--watch- watch for changes and rebuild
- OPENAI_API_KEY - OpenAI API key (default:
null) - JENNGEN_CACHE - cache folder (default:
.jenngen_cache) - JENNGEN_DIST - output folder (default:
.dist) - JENNGEN_INSTRUCTIONS - instructions file (default:
.jenngen) - JENNGEN_LIVERELOAD_PORT - livereload port (default:
35729) - JENNGEN_MODEL - OpenAI model (default:
gpt-4-1106-preview) - JENNGEN_PORT - server port (default:
3000) - JENNGEN_OLLAMA_MODEL - Use a local Ollama model; overrides OPENAI_API_KEY (default:
null)
The ./website folder is generated and deployed to https://thrashr888.github.io/jenngen/.
Your index.html file:
hello world
The generated index.html file:
<!doctype html>
<html>
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Hello World</title>
<script src="https://cdn.tailwindcss.com"></script>
</head>
<body>
<h1 class="text-3xl font-bold underline">
Hello world!
</h1>
</body>
</html>
"Welcome to the JennGen website!"
{title}
[beautiful hero section explaining the product]
[insert high quality marketing content here]
## Getting Started
Using JennGen is easy. Just follow these steps:
1. Get your OpenAI key from [https://beta.openai.com/](https://beta.openai.com/)
2. Use npx to get jenngen and render the current folder: `npx jenngen .`
3. Open [http://localhost:3000](http://localhost:3000) to view it in the browser.
... insert more of a product story here ...
[insert a centered cat picture here from placekitten.com]
<time current>
<sparkles animated="true" count="3">
[footer with links to about, privacy, and terms]
See the ./website folder and https://thrashr888.github.io/jenngen/ for more examples.
- Add quality examples for more languages and file formats.
- Add a way to use a local model with Ollama.
- Add RAG with local markdown docs.
- I'm pretty sure there are bugs with npx and the JennGen example folders because the cwd confuses me.
- Add a way to serve the output folder without watching. I think we're double-watching and reloading.
- Use JennGen to generate a website for its GitHub Pages.
- Record a demo video for the readme, website, and Twitter.
- Support website layouts. Maybe a
_layout.htmlfile?