Intelligent fill (AI-powered)

Skip the manual data-mapping step entirely: send raw, unstructured text (an email, a transcript, free-form notes) alongside your docx template, and docMaker's LLM extracts the structured fields and fills the document for you.

Unlike Create pdf from docx template, which requires you to send pre-structured JSON data matching your template's tags, intelligent_fill accepts:

• rawText — the unstructured text to parse (an email body, meeting notes, a form submission, etc.)

• userPrompt (optional) — extra instructions telling the LLM how to interpret the raw text or handle edge cases

docMaker reads your template's tags, asks the LLM to extract matching values from rawText, validates the result against the template's schema, then fills and converts the document exactly like docx_fill_convert — returning a PDF or DOCX.

It accepts every parameter from docx_fill_convert (webhook, email delivery, Bubble delivery, output format, etc.) in addition to rawText and userPrompt.

Don't have a docx template yet? See Build your docx template to learn how to create one — intelligent_fill uses the exact same templates as docx_fill_convert, just with the LLM filling in the tags instead of you sending structured JSON.

Async, like every docMaker document route: the call returns 202 with a jobId immediately. Poll GET /jobs/{jobId} or set a webhook_url to get the result.

Credits: this route costs more credits per call than a standard template fill, since it includes an LLM extraction step in addition to document generation.

Click "Next" to see the full list of intelligent_fill API parameters.