Skip to main content
The Word endpoint produces professional .docx output with proper heading hierarchy, line spacing, tables, and (optionally) images. The agent auto- accepts the inline preview diff on the backend, so edits apply immediately — no user-in-the-loop required.

​
What the Word agent can do

Ordered roughly from “used in almost every document” to specialized:
  • Paragraph and run styling — heading levels, body styles, custom paragraph styles; bold, italic, underline, strikethrough, double strikethrough, superscript, subscript, all-caps, small-caps, hidden, highlight color, shading color, custom font + size + color; per-run hyperlinks; line spacing, space-before / space-after, indents, alignment (left / center / right / justify).
  • Lists — bulleted and numbered, with custom markers and nesting levels; per-paragraph indent + marker control for designs like • … / – … / 1. ….
  • Tables — headered tables with shaded header row, per-cell borders and shading, cell merge / split, column-width control, alignment per cell, text formatting inside cells.
  • Document structure — headers, footers (with left/center/right sections), section breaks (continuous, next-page, even/odd), column breaks, multi-column page layouts, drop caps, watermarks, page numbering fields.
  • Media — images, charts (same chart types as Excel), shapes, text boxes (linked or standalone).
  • Citations as hyperlinks — inline links to sources; pass context_files to give the agent specific references to cite.
  • Footnotes and endnotes — full per-note content, with automatic renumbering.
  • Review / collaboration — comments (with threaded replies and resolve), tracked-changes (accept / reject, individually or all at once).
  • References — cross-references to headings, bookmarks, footnotes, tables, figures; merge fields for mail-merge flows; form controls (rich text, plain text, drop-down, checkbox); sub-documents and master-doc structure; document compare.
  • Equations / math — full Equation schema for fractions, radicals, limits, integrals, sums, matrices, Greek letters. Renders natively in the embedded viewer and in any downloaded .docx.
  • Styles catalog — create new paragraph / character styles, modify or delete existing ones, set a document-wide default style.
  • Page fields — date, time, page number, total pages, file name, author; auto-updated on open.
Describe the document’s purpose, audience, and tone in your prompt; the agent picks the right styles, layouts, and structure. You can steer with explicit hints (“use two-column layout for the analysis section”, “start Chapter 1 with a drop cap”, “legal / contractual tone”) and the agent respects them.

​
Quickstart

resp = requests.post(
    f"{API}/word/generate",
    headers={"X-API-Key": KEY},
    json={
        "prompt": (
            "Write a formal acceptance letter for an engineering manager role "
            "at Acme Corp dated April 17, 2026. Signed by Jane Smith."
        ),
    },
)
Output: a 3+ paragraph letter with header/footer, proper typography, date right-aligned, sign-off block.

​
Common patterns

​
Reports

prompt = """
Produce a 4-page Q3 business review report for Acme SaaS:
- Executive summary (2 paragraphs)
- Revenue & growth (analysis paragraphs, one bullet list)
- Product highlights (3 subsections)
- Q4 priorities

Tone: professional, concise. Use Heading 1/2 hierarchy.
"""
resp = requests.post(f"{API}/word/generate", headers=headers,
                     json={"prompt": prompt})
The agent writes in sections with clean heading hierarchy, 1.15 line spacing, and an accent color on H1s.

​
Letters and memos

prompt = (
    "Write a formal termination letter. "
    "Recipient: Mr. John Doe, Senior Consultant, Acme LLC, 123 Broadway, NY 10001. "
    "Reason: ending of contract at end of current term (June 30, 2026). "
    "Tone: polite, firm. "
    "Include sign-off for Jane Smith, CEO."
)

​
Editing an existing document

# 1. Upload
upload = requests.post(
    f"{API}/files",
    headers=headers,
    files={"file": open("draft_contract.docx", "rb")},
    data={"purpose": "init"},
).json()

# 2. Edit
resp = requests.post(f"{API}/word/generate", headers=headers, json={
    "prompt": "Replace the payment terms section with net-30 payment terms.",
    "init_file": upload["file_id"],
})
The agent:
  1. Inspects the document to find the existing payment-terms section
  2. Generates a PreviewActionPlan with the replacement
  3. Auto-accepts the diff (no user clicking needed)
  4. Returns the updated .docx

​
Working with images

Upload images first, then reference them by image_id: 
hero = requests.post(
    f"{API}/images",
    headers=headers,
    files={"file": open("logo.png", "rb")},
    data={"tags": '["logo"]'},
).json()

resp = requests.post(f"{API}/word/generate", headers=headers, json={
    "prompt": "Draft a company prospectus with our logo at the top",
    "image_assets": [{"image_id": hero["image_id"], "tags": ["logo"]}],
})
If you reference “logo” or “hero image” in the prompt but don’t attach images, the preflight validator will reject the request upfront.

​
Writing quality

The underlying agent is tuned for professional prose. It avoids:
  • Banned words: delve, leverage, utilize, robust, seamless, cutting-edge, etc.
  • Banned transitions: Furthermore, Moreover, Additionally, In summary…
  • Banned phrases: “in today’s fast-paced world”, “at the end of the day”
  • Em dashes as a stylistic tic (max one per page)
The output reads like senior human professional work. If you want a specific tone, specify it: “formal academic”, “conversational but direct”, “legal/contractual”. The agent obeys tone instructions.

​
Things to watch for

  • Table of contents is not generated today. If your prompt asks for a TOC, the agent skips it and notes the omission in summary.
  • Long documents (>5 pages) are built page-by-page. Specify “first page preview only” if you want a quick iteration.