read_file
Read document content from a DOCX file or Google Doc. Output is token-limited to approximately 14k tokens by default and includes pagination metadata (has_more, next_offset). Use offset and limit to page through long documents.
The default
toon format is compact and agent-friendly. It renders inline formatting as short tags — <b>bold</b>, <i>italic</i>, <u>underline</u>, <highlighting>highlighted</highlighting>, <a href="...">link</a> — so you can see both content and structure without parsing raw XML. Switch to json for structured output or simple for plain paragraph text.Path to the DOCX file. Provide either
file_path or google_doc_id.Google Doc ID or URL. Extract the ID from the URL:
docs.google.com/document/d/{ID}/edit.1-based paragraph offset for pagination. Negative values count from the end of the document.
Maximum number of paragraphs to return. When omitted, output is token-limited to approximately 14k tokens and includes
has_more and next_offset in the response.Fetch specific paragraphs by their
_bk_* IDs instead of a contiguous range.Output format. One of
toon (default, compact with inline tags), json (structured), or simple (plain text, no tags).When
true (default), includes inline formatting tags (<b>, <i>, <u>, <highlighting>, <a>). Set to false to emit plain text with no inline tags.Pagination example
When the response contains"has_more": true, pass the returned next_offset value as the offset parameter in your next call to continue reading.
grep
Search paragraphs using regular expressions. Use file_path for session-based search, file_paths for stateless multi-file search, or google_doc_id for Google Docs.
Hints: readOnly — does not modify any documents.
Path to the DOCX file for session-based search.
Google Doc ID or URL. Extract the ID from the URL:
docs.google.com/document/d/{ID}/edit.Multiple file paths for stateless multi-file search. No session is created when using this parameter.
A single regex pattern to search for.
Multiple regex patterns. Matches any paragraph that satisfies at least one pattern.
When
true, the search is case-sensitive. Defaults to false.When
true, matches only whole words.Maximum number of results to return.
Number of surrounding characters to include in each result snippet.
When
true, returns at most one result per paragraph even if multiple matches exist.When
true, searches the raw XML (word/document.xml) instead of paragraph text. Useful for inspecting markup.When
false, skips document view context such as list labels and headers, giving faster results. Defaults to true.has_tracked_changes
Check whether the document body contains tracked-change markers, including insertions, deletions, moves, and property-change records.
Hints: readOnly — does not modify the document.
Path to the DOCX file.
get_file_status
Return file and session metadata for an open document, including edit count, normalization stats, and cache information.
Hints: readOnly — does not modify any documents.
Path to the DOCX file. Provide either
file_path or google_doc_id.Google Doc ID or URL. Extract the ID from the URL:
docs.google.com/document/d/{ID}/edit.