Every enterprise that processes documents eventually hits the same wall. You can extract data from invoices, claims, financial filings, and medical records. But the moment you need to put data back into a PDF, things fall apart.
This is the form filling problem, and it is far more difficult than most teams realize.
Pulse Form Fill builds on top of our state-of-the-art extraction models, leveraging the same visual document understanding that powers Pulse's extraction pipeline as the foundation for accurately placing data back into any PDF.
The real challenge is not fillable PDFs
If every PDF had clean, embedded form fields, form filling would be straightforward. But in practice, most enterprise PDFs do not have defined form fields at all. Government forms arrive as scanned images. Insurance claim documents are flat PDFs with no interactive elements. Healthcare intake forms are printed, photocopied, and re-digitized until any trace of structured metadata is gone.
When there are no checkboxes to check, no text fields to populate, and no dropdown menus to select from, you need a system that can visually understand the document, identify where data should go, and place it with precision.
That is exactly what Pulse Form Fill does.
How Pulse Form Fill works
Pulse Form Fill is a set of API endpoints that let you programmatically edit and fill any PDF.
Under the hood, Pulse uses its vision-language models to analyze the full document layout, understanding labels, lines, boxes, whitespace patterns, and spatial relationships between elements. The system detects where fillable regions exist based on visual context and inserts values in the correct locations. Checkboxes that are just empty squares on a scanned page get checked. Text that belongs next to a label gets placed precisely. Tables get populated row by row.
You provide structured data or natural language instructions. Pulse returns a filled PDF.

Beyond filling, Form Fill also supports clearing existing field values, allowing you to reset forms programmatically before re-populating them with updated data. This makes it easy to build repeatable, automated pipelines around documents that need to be revised or reprocessed.
Where healthcare teams are using this today
Healthcare is one of the industries where this problem is most acute. Pre-authorization forms, CMS-1500 claims, patient intake documents, and provider enrollment packets all require data to flow from one system into a specific PDF format. These forms change across payers, regions, and plan types, making it impractical to maintain static templates for every variant.
Healthcare companies using Pulse Form Fill are automating workflows that previously required staff to manually open each PDF, locate the right fields by eye, and type values one by one. With Pulse, the same workflow runs through a single API call. Data extracted from clinical notes or EHR systems gets routed directly into the target form, filled accurately, and returned as a completed document.
Built for scale and speed
Our research team has invested significant effort into making Form Fill both accurate and fast at production volumes. The underlying models are optimized for low latency so that form filling can run inline within automated workflows, not as a batch job you check on later. Whether you are filling ten forms or ten thousand, the system scales without manual intervention.
End-to-end document pipelines
Form Fill is designed to work alongside the rest of the Pulse platform, giving you the building blocks to construct complete document automation pipelines. A typical workflow might look like this:
- Split a multi-document PDF into individual forms using Pulse Split.
- Extract structured data from source documents using Pulse Extract with a defined schema.
- Fill target forms with the extracted data using Pulse Form Fill.
- Clear and re-fill forms as upstream data changes, keeping documents in sync without manual rework.
Each step is a standalone API call, which means you can compose pipelines that match your specific workflow rather than conforming to a rigid, one-size-fits-all system.
If your team is automating document workflows and form filling is the bottleneck, we would like to hear from you here.
