docx.sh — DOCX to HTML

docx.sh exposes the same engine over MCP and REST. Authenticate with a Bearer token — generate one in Agents after signing up.

Authentication

Header: Authorization: Bearer <token>
REST base URL: https://docx.sh/api/v1
MCP endpoint: https://docx.sh/mcp

Reference

MCP tools map 1:1 to REST endpoints. Pick the transport that fits your agent or backend.

Connect via Agents setup (Cursor, Claude Code, Codex). Each tool call counts toward your monthly MCP/API limit except HTML/DOCX downloads.

Example MCP config

Project-level config — treat tokens like secrets.

{
  "mcpServers": {
    "docxtohtml": {
      "url": "https://docx.sh/mcp",
      "headers": {
        "Authorization": "Bearer <token>"
      }
    }
  }
}

`upload_docx`

Upload a DOCX file for conversion or modification.

Params: filename (string), content_base64 (string)

Returns: Document metadata with id and status.

`convert_docx`

Convert an uploaded document to HTML preview and full HTML.

Params: document_id (string)

Returns: Updated document with preview_path and html_download_path.

`modify_docx`

Apply structured element operations and return a new derived document as DOCX or HTML.

Params: document_id (string), output_format (docx|html), operations (array of replace, insert_paragraph, replace_paragraph, delete_paragraph, set_cell_text, add_comment, …)

Returns: New child document; source document stays unchanged.

`get_docx_content`

Inspect elements for editing: paragraph or run anchors, table coordinates, optional comments and tracked changes.

Params: document_id (string), at? (string), limit? (integer), offset? (integer), mode? (paragraphs|runs), include? (comments, revisions)

Returns: Paginated lines with path, wt:N, kind, table coords; optional comments/revisions.

`list_documents`

List your active (non-expired) documents.

Params: —

Returns: Array of document metadata.

`get_document`

Get document metadata and authorized download paths.

Params: document_id (string), include_preview? (boolean)

Returns: Full document object.

`get_html`

Get HTML metadata for a document (use API paths to download bodies).

Params: document_id (string), format? (preview|html|both)

Returns: Paths and sizes for preview and/or full HTML.

`delete_document`

Delete a document and its generated files.

Params: document_id (string)

Returns: Confirmation with deleted id.

HTTP JSON API for scripts, CI, and backends.

Upload and convert

Replace DOCUMENT_ID with the id from the upload response.

# Upload
curl -X POST "https://docx.sh/api/v1/documents" \
  -H "Authorization: Bearer <token>" \
  -F "file=@document.docx"

# Convert to HTML
curl -X POST "https://docx.sh/api/v1/documents/DOCUMENT_ID/convert" \
  -H "Authorization: Bearer <token>"

# Download HTML
curl "https://docx.sh/api/v1/documents/DOCUMENT_ID/html" \
  -H "Authorization: Bearer <token>" -o output.html

Method	Path	Description
GET	`/api/v1/documents`	List active documents.
POST	`/api/v1/documents`	Upload a DOCX file (multipart form field: file).
GET	`/api/v1/documents/:id`	Get document metadata.
DELETE	`/api/v1/documents/:id`	Delete a document.
POST	`/api/v1/documents/:id/convert`	Convert document to HTML.
POST	`/api/v1/documents/:id/modify`	Modify document; returns a new derived document.
GET	`/api/v1/documents/:id/content`	Inspect elements (mode=paragraphs\|runs, include=comments,revisions).
GET	`/api/v1/documents/:id/preview`	Download first-page HTML preview.
GET	`/api/v1/documents/:id/html`	Download full HTML (?inline=true for inline).
GET	`/api/v1/documents/:id/zip`	Download HTML bundle as ZIP.
GET	`/api/v1/documents/:id/docx`	Download modified DOCX output.

Developers