Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/UnkleFunk/HouseMusicSwarm-/llms.txt

Use this file to discover all available pages before exploring further.

The Docs Agent is OpenSwarm’s document engineer. It creates professional Word documents (.docx) and PDFs from HTML or Markdown content, edits existing documents with surgical search-and-replace precision, converts between formats, and maintains automatic version history for every DOCX export. HTML is used as the canonical source format because it gives the agent complete control over fonts, colors, tables, spacing, and layout — all of which survive conversion to DOCX and PDF with high fidelity.

​
Tools

Creates a new document from HTML or Markdown content.
  • HTML workflow → produces a .source.html file (the canonical source of truth). Always auto-converts to .docx immediately after creation.
  • Markdown workflow → produces a .md file only. No .docx or .pdf is generated from Markdown.
Key parameters:
  • project_name — folder name under ./mnt/ (e.g., "quarterly_reports")
  • document_name — filename without extension (e.g., "q3_executive_summary")
  • content — {"type": "html", "value": "<!DOCTYPE html>..."} or {"type": "markdown", "value": "# Title\n..."}
CreateDocument(
    project_name="client_proposals",
    document_name="acme_proposal",
    content={"type": "html", "value": "<!DOCTYPE html><html>..."}
)
Edits an existing document’s HTML source. Two modes:search_and_replace (preferred): provide a list of {old_content, new_content} pairs. Each old_content is matched exactly — like StrReplace. Batch all changes into a single call. The file is not modified if any replacement fails.
ModifyDocument(
    project_name="client_proposals",
    document_name="acme_proposal",
    operation="search_and_replace",
    replacements=[
        {"old_content": "#C8102E", "new_content": "#DA291C"},
        {"old_content": "<h1>Draft Proposal</h1>", "new_content": "<h1>Final Proposal</h1>"}
    ]
)
Line operations: replace, insert, or delete by line number — for structural additions or deletions where no existing content can be matched.
ModifyDocument only updates the HTML source. Call ConvertDocument to re-export to DOCX or PDF after editing.
Reads the current HTML source of a document. Always run this before editing to see the exact content and line numbers.Supports an optional view_range parameter (e.g., [1, 50]) to show only a portion of a large document.
Lists all documents in a project folder. Shows:
  • .source.html files (canonical source)
  • All DOCX exports including versioned files (report.docx, report_v2.docx, report_v3.docx)
  • Whether each DOCX has a companion snapshot for version restore
  • Other converted formats (PDF, Markdown, TXT) and their sizes
Converts the .source.html to another format. Supported output formats:
FormatNotes
docxWord document. DOCX exports are automatically versioned — existing files are never overwritten; each export becomes report.docx, report_v2.docx, etc. A companion .snapshot.html is saved alongside every DOCX for version history.
pdfHigh-quality PDF rendered with WeasyPrint. Unicode normalization is applied automatically.
markdownMarkdown, useful for documentation sites and wikis.
txtPlain text — all formatting stripped.
Restores the working .source.html to the state it was in at a previous DOCX export.Every ConvertDocument call that produces a .docx automatically saves a companion snapshot:
report.docx
report.docx.snapshot.html      ← HTML source at time of that export
report_v2.docx
report_v2.docx.snapshot.html
RestoreDocument reads the named snapshot and writes it back as the canonical .source.html, ready for further edits or re-conversion. Use ListDocuments to see all available versions.
RestoreDocument(
    project_name="client_proposals",
    docx_filename="acme_proposal_v2.docx"
)

​
Supported formats

FormatReadWriteNotes
HTML✅ (source input)✅ (.source.html)Canonical source format; full CSS styling control
DOCX (Word)—✅Auto-versioned; companion snapshot saved on every export
PDF—✅High-quality via WeasyPrint; unicode normalization applied
Markdown✅ (source input)✅ (.md)Markdown workflow does not produce DOCX or PDF
TXT—✅Plain text, all formatting stripped

​
Output files

Every file the Docs Agent creates or exports is saved under ./mnt/<project_name>/documents/ and the full path is included in the agent’s response. Use the path to open the file, share it, or hand it off to another agent. 
./mnt/client_proposals/documents/acme_proposal.source.html
./mnt/client_proposals/documents/acme_proposal.docx
./mnt/client_proposals/documents/acme_proposal.docx.snapshot.html
./mnt/client_proposals/documents/acme_proposal.pdf
If you provide an output path outside the project folder, the agent creates the file in the project folder first, then copies it to your specified location with CopyFile.
RestoreDocument gives you full version history. If you export a DOCX, keep editing, and then want to roll back to the version from two exports ago, just call RestoreDocument with the earlier DOCX filename. The snapshot is always there.

​
Example prompts

Write an executive summary for my Series A pitch. Include our traction metrics and a competitive landscape section.

Convert this HTML file to a Word document. Keep the formatting as close as possible.

Create a research report with these findings and export it as a PDF ready for sharing.

I need a one-page onboarding SOP for a remote operations coordinator. Deliver it as Markdown.

Edit the proposal I just created — change the accent color from #C8102E to #DA291C and update the title to "Final Proposal".

Build docs developers (and LLMs) love

Get started for freeTalk to us