Batch Vision

This is a tool for taking a bunch of images and getting specific pieces of data out of each one. I used it to take hundreds of receipts and get the date and total from each. LLMs are quite good at reading handwriting and figuring out which number in the image is the one you want.

If you notice any bugs or have any feature requests, please open an issue on GitHub.

WarningLLMs don't always get it right. I was able to get my error rate down to 3% by running my data twice and checking differences between the runs, but your results may vary. Exercise an amount of caution appropriate to the seriousness of any mistakes.

This tool does not collect any revenue. Instead, you will need to set up an account with OpenAI and purchase credit with them. Create an API key and paste it into the box below; that will charge your usage of this tool to your account with OpenAI.

LLM used: GPT-4o. Pricing is roughly 2 cents per megabyte of images. You can track your usage here.

This will only be used for extracting data from the images you upload. We don't keep a copy or use it for anything else.

Generated Schema

{
  "type": "object",
  "properties": {
    "date": {
      "type": "string",
      "description": "The date of the receipt in mm/dd/yyyy format."
    },
    "amount": {
      "type": "number",
      "description": "The amount, including tips, in monetary format."
    }
  },
  "required": [
    "date",
    "amount"
  ],
  "additionalProperties": false
}

Generated Prompt

Drag and drop files here

File NameStatusdateamount