Fill in PDF forms with AI
Give the agent a form and its answers. It can inspect interactive fields, position text on a flat page and produce a separate file for checking.
To fill in PDF forms with AI, ask the agent to inspect the document before you give it the answers. With the pdf skill, it can work out whether the file has interactive fields or needs text placed on the page. That decision matters when a form has several similar boxes or a signature area that must stay clear.
Consider a delivery form with a reference number and two dates. A fillable file has field identifiers the agent can inspect. On a flat version, a page image helps check that the reference sits on the intended line and the text fits the available space.
What it does
- Inspect field identifiers and types in an interactive form.
- Enter supplied answers in a separate copy.
- Place text on a flat form using page coordinates.
- Review an image of a completed page for overlap.
- Extract selected pages when only part of the form is needed.
Install and use
- Install the skill in a file-capable agent with Python and the PDF libraries for your task. Image-based checks need a PDF renderer.
- Ask “Fill the order number and delivery date in a copy. Leave the signature box empty.”
- Open the result and compare each answer with its label before sending the form.
FAQ
Source
Made by Anthropic. Original skill: pdf, published on GitHub under the Apache-2.0 licence.
View on GitHub