Google Productivity - mAIvn Docs

#Google productivity

GoogleDocsToolSet, GoogleSheetsToolSet, and GoogleSlidesToolSet wrap the Docs, Sheets, and Slides APIs respectively. They share the OAuth bearer token handling from google_workspace._shared.

All three toolsets are content-level editors -- they assume you already have the document/spreadsheet/presentation ID. To find one by name, use GoogleDriveToolSet.search_files(include_ids=True) and pass the returned file dict into any of these tools. The Docs / Sheets / Slides tools accept either a raw ID string or a Drive file dict (or a dict from another tool in the same toolset) and will pull the right ID out.

#GoogleDocsToolSet

Python

from maivn_tools import GoogleDocsToolSet

connector = GoogleDocsToolSet(token=token_cache)

Tool	Permission	Description
`get_document(document_id)`	READ	Fetch a Doc. Accepts a raw doc ID or a Drive file dict.
`create_document(title)`	WRITE	Create a blank Doc.
`batch_update(document_id, requests, write_control)`	WRITE	Apply Docs API update requests.
`insert_text(document_id, text, index)`	WRITE	Insert text.
`replace_text(document_id, find, replace, match_case)`	WRITE	Find/replace.
`delete_content_range(document_id, start_index, end_index)`	DELETE (destructive)	Delete a range.

#Agent-ready behavior

GoogleDocsToolSet has no list/search of its own -- use GoogleDriveToolSet.search_files(include_ids=True) to locate a Doc, then pass the returned file dict to get_document / batch_update / insert_text / replace_text / delete_content_range. The connector reads doc_id / document_id / documentId / file_id / id out of the dict.
create_document returns the new Doc resource; pass its documentId (or the whole dict) to the edit tools to populate it.

Python

# Find then edit, without unwrapping IDs by hand.
file = drive.search_files("name = 'Launch plan'", include_ids=True)["files"][0]
docs.insert_text(file, text="Status update\n")

#Destructive tools

delete_content_range is tagged destructive and requires PermissionFlag.DELETE. Filter it off the toolset with exclude_tags=["destructive"] to exclude tagged destructive methods.

#GoogleSheetsToolSet

Python

from maivn_tools import GoogleSheetsToolSet

connector = GoogleSheetsToolSet(token=token_cache)

Tool	Permission	Description
`get_spreadsheet(spreadsheet_id, include_grid_data, ranges)`	READ	Metadata + grid data. Accepts a raw ID or a Drive file dict.
`create_spreadsheet(title, sheets)`	WRITE	Create a spreadsheet.
`get_values(spreadsheet_id, range, major_dimension, value_render_option, date_time_render_option)`	READ	Read a range.
`batch_get_values(spreadsheet_id, ranges, major_dimension, value_render_option)`	READ	Multi-range read.
`update_values(spreadsheet_id, range, values, value_input_option)`	WRITE	Overwrite a range.
`append_values(spreadsheet_id, range, values, value_input_option, insert_data_option)`	WRITE	Append rows.
`clear_values(spreadsheet_id, range)`	DELETE (destructive)	Clear a range.
`batch_update(spreadsheet_id, requests)`	WRITE	Spreadsheet-level updates.
`add_sheet(spreadsheet_id, title, row_count, column_count)`	WRITE	Add a sheet.
`delete_sheet(spreadsheet_id, sheet_id)`	DELETE (destructive)	Remove a sheet.

#Agent-ready behavior

GoogleSheetsToolSet is a value/range editor -- use GoogleDriveToolSet.search_files(include_ids=True) to locate a spreadsheet, then pass the returned file dict to any of these tools. The connector reads sheet_ref_id / spreadsheet_id / spreadsheetId / file_id / id out of the dict.
get_spreadsheet returns the full Sheets resource; with include_grid_data=True and ranges you also get cell-level values.
create_spreadsheet returns the new resource; pass its spreadsheetId (or the whole dict) to the value / batch_update tools.

Python

file = drive.search_files("name = 'Q4 forecast'", include_ids=True)["files"][0]
sheets.get_values(file, range="Plan!A1:D20")

#Destructive tools

clear_values and delete_sheet are tagged destructive and require PermissionFlag.DELETE. clear_values removes cell contents but leaves formatting; delete_sheet removes an entire tab. Filter them off the toolset with exclude_tags=["destructive"] to exclude tagged destructive methods.

#GoogleSlidesToolSet

Python

from maivn_tools import GoogleSlidesToolSet

connector = GoogleSlidesToolSet(token=token_cache)

Tool	Permission	Description
`get_presentation(presentation_id)`	READ	Fetch presentation. Accepts a raw ID or a Drive file dict.
`get_page(presentation_id, page_object_id)`	READ	One slide / master / layout.
`get_page_thumbnail(presentation_id, page_object_id, thumbnail_size, mime_type)`	READ	Slide thumbnail URL.
`create_presentation(title)`	WRITE	New deck.
`batch_update(presentation_id, requests)`	WRITE	Slides API update requests.
`create_slide(presentation_id, insertion_index, layout, object_id)`	WRITE	Insert a slide.
`insert_text(presentation_id, object_id, text, insertion_index)`	WRITE	Add text to a shape.
`replace_all_text(presentation_id, find, replace, match_case)`	WRITE	Find/replace.
`delete_object(presentation_id, object_id)`	DELETE (destructive)	Remove a slide / shape.

#Agent-ready behavior

GoogleSlidesToolSet is a slide/shape/text editor -- use GoogleDriveToolSet.search_files(include_ids=True) to locate a deck, then pass the returned file dict to any of these tools. The connector reads presentation_id / presentationId / file_id / id out of the dict.
get_presentation returns the full Slides resource with slides[*], masters[*], layouts[*]. Each slide's objectId is the page_object_id argument for get_page / get_page_thumbnail, and the input for the object_id argument of insert_text / delete_object.
create_presentation returns the new resource; pass its presentationId (or the whole dict) to the edit tools.

Python

file = drive.search_files("name = 'All-hands deck'", include_ids=True)["files"][0]
deck = slides.get_presentation(file)
slides.replace_all_text(file, find="{{date}}", replace="2026-05-16")

#Destructive tools

delete_object is tagged destructive and requires PermissionFlag.DELETE. It removes a slide, shape, or other page object permanently. Filter it off the toolset with exclude_tags=["destructive"] to exclude tagged destructive methods.