Transloadit's 213-line llms.txt shows what thorough AI preparation looks like

# Transloadit

> Transloadit is a versatile file uploading and processing API that allows developers to create complex media processing workflows through declarative JSON recipes called Assembly Instructions.

## Uploading and importing (INPUT)

Transloadit handles file uploads over XHR or tus.io (our open source protocol for resumable file uploads) and can process files through various "Robots" (specialized processing steps) that can be chained together to create customized workflows.

## Encoding workflows (PROCESS)

Assembly Instructions at a glance:

```json
{
  "steps": {
    ":original": {
      "robot": "/upload/handle"
    },
    "browser720_webm_encoded": {
      "use": ":original",
      "robot": "/video/encode",
      "preset": "webm",
      "width": 1280,
      "height": 720
    },
    "browser720_h264_encoded": {
      "use": ":original",
      "robot": "/video/encode",
      "preset": "mp4",
      "width": 1280,
      "height": 720
    },
    "thumbed": {
      "use": "browser720_h264_encoded",
      "robot": "/video/thumbs",
      "width": 1280,
      "height": 720,
      "format": "jpg",
      "resize_strategy": "fit",
      "count": 1,
    },
    "exported": {
      "use": ["browser720_webm_encoded", "browser720_h264_encoded", "thumbed", ":original"],
      "robot": "/s3/store",
      "credentials": "demo_s3_credentials",
      "url_prefix": "https://demos.transloadit.com/"
    }
  }
}
```

As you can see, our features are called Robots, and each Step inside
the Assembly Instructions can refer one. Robots can pass files to each other
with the `use` keyword. This means you can create complex workflows
unique to your application.

## Exporting to cloud storage or downloading results (OUTPUT)

Transloadit can export to all the major cloud storage providers.
Customers first need to store their credentials in their Transloadit account and
then refer to them in the Assembly Instructions.

Alternatively, Transloadit keeping 24h of temporary storage, you can directly download
resulting files when an Assembly is finished.

All meta data is available on a unique Assembly URL which contains the Assembly Status JSON.

## Example integration code with Node.js

```js
// npm install transloadit
const { Transloadit } = require('transloadit')

const transloadit = new Transloadit({
  authKey: 'TRANSLOADIT_KEY',
  authSecret: 'TRANSLOADIT_SECRET',
})

const options = {
  files: {
    file1: '/PATH/TO/FILE.jpg',
  },
  params: {
    steps: {
      // You can have many Steps. In this case we will just resize any inputs (:original)
      ':original': {
        robot: '/upload/handle',
      },
      resize: {
        use: ':original',
        robot: '/image/resize',
        result: true,
        width: 75,
        height: 75,
      },
    },
    // OR if you already created a template, you can use it instead of "steps":
    // template_id: 'YOUR_TEMPLATE_ID',
  },
  waitForCompletion: true, // Wait for the Assembly (job) to finish executing before returning