Skip to content

jliebster/make-real

 
 

Repository files navigation

Make Real

Use this repo as a template to create Make Real style apps like makereal.tldraw.com. To get started:

  1. Use the template and clone your new repo to your computer
  2. Run npm install to install dependencies
  3. Get an OpenAI API key from platform.openai.com/api-keys. Make sure you are at least a Tier 1 API user, which means you have access to GPT-4 Vision. You can check your tier on the OpenAI API Limits.
  4. Create a .env.local file that contains OPENAI_API_KEY=your api key here
  5. Run npm run dev
  6. Open localhost:3000 and make some stuff real!

How it works

Make Real is built with tldraw, a very good React library for creating whiteboards and other infinite canvas experiences.

To use it, first draw a mockup for a piece of UI. When you're ready, select the drawing, and press the Make Real button. We'll capture an image of your selection, and send it to OpenAI's GPT-4V along with instructions to turn it into a HTML file.

We take the HTML response and add it to a tldraw custom shape. The custom shape shows the response in an iframe so that you can interact with it on the canvas. If you want to iterate on the response, annotate the iframe, select it all, and press 'Make Real' again.

To make changes

To change how Make Real works, start from the makeReal() function. From there, you can change the prompt that gets sent to gpt-4.

If you'd like Make Real to create something other than HTML, you'll need to either update the PreviewShape to display something different, or use one of tldraw's built-in shapes like image or text.

The dangerous API key input method

For prototyping or at least until the vision APIs are able to support higher usage limits, we've also included the RiskyButCoolAPIKeyInput, similar to the one found on makereal.tldraw.com. Please use this as carefully and ethically as you can, as users should be reluctant to add API keys to public sites.

Releases

No releases published

Packages

No packages published

Languages

  • TypeScript 92.2%
  • CSS 7.5%
  • JavaScript 0.3%