Upload the image you wish to read text from to your S3 bucket. I used a screenshot from a book in my Kindle library.

Goto IAM and assign AmazonTextractFullAccess to your user.
On your host machine that has node installed, create a directory called textract.
Create a package.json file in that directory:
{
"name": "textract",
"version": "1.0.0",
"description": "Read text from an image",
"dependencies": {
"aws-sdk": "^2.1053.0"
}
}
Create an index.js file and put this code in it replacing YOUR_ values:
const fs = require('fs')
const AWS = require("aws-sdk")
const awsObj = new AWS.Textract({
accessKeyId: YOUR_ACCESS_KEY_ID,
secretAccessKey: YOUR_SECRET_ACCESS_KEY,
region:YOUR_BUCKET_REGION
});
let file = 'YOUR_screenshot.png'
let params = {
Document: {
S3Object: {
Bucket: 'YOUR_BUCKETNAME,
Name: file
}
},
FeatureTypes: ["TABLES"]
}
awsObj.analyzeDocument(params, (err, data) => {
if (err) {
console.log(err)
} else {
//console.log(data)
let text = ""
for(let i in data.Blocks) {
if (data.Blocks[i].BlockType !== "LINE") {
continue
}
console.log(data.Blocks[i])
// if the line starts at a point from the Left that is greater than .01, assume the line is indented and start a new line
if (data.Blocks[i].Geometry.BoundingBox.Left > .01) {
text+="\n"
}
text+=" " + data.Blocks[i].Text
}
console.log(text)
fs.writeFile(file + ".txt", text, function (err) {
if (err) return console.log(err);
console.log('done');
});
}
})
In the textract directory, on the command line run:
npm install
And then run:
node index.js
You'll get a lot of interesting output and the text in the image written to a file on your host machine as well as printed to the console.