Create a new, empty Google Slides presentation. (1 API call.)

Args: title: The title for the new presentation.

Returns: {presentationId, title, url}.

Get a presentation as a bounded, structured overview. (1 API call.)

By default returns a trimmed summary (slides, elements, positions, short text snippets) so large decks stay within a finite output budget.

Args: presentation_id: The presentation ID. include_masters: Also summarize layouts and masters. raw: Return the full, untrimmed API response instead of the summary.

Get a single page (slide) as a bounded summary. (1 API call.)

Args: presentation_id: The presentation ID. page_id: The page (slide) object ID. raw: Return the full, untrimmed API response instead of the summary.

List slides with index, id, and layout reference only. (1 API call.)

The lightest-weight read — use it to find slide IDs before deeper calls.

Apply raw Slides batchUpdate requests atomically. (1 API call.)

This is the low-level power tool: requests is a list of Slides API request objects (e.g. createSlide, insertText, createShape, updateTextStyle, updatePageElementTransform, duplicateObject ...). See https://developers.google.com/workspace/slides/api/reference/rest/v1/presentations/request Many edits batched into one call still cost a single write quota unit.

Note: updateShapeProperties/updateTextStyle/updatePageProperties require a fields mask; a missing mask is reported as a warning.

Args: presentation_id: The presentation ID. requests: A list of single-key Slides request objects.

Returns: {replies, warnings} where replies is the API response array.

Set an element's position (and optionally size) in points. (Up to 2 calls.)

A convenience over updatePageElementTransform that lets you place a bounding box by its top-left corner in points without computing an affine matrix. Reads the element's current transform once (to preserve scale/shear when size is omitted), then applies one absolute transform update.

Args: presentation_id: The presentation ID. element_id: The page element to move/resize. x_pt: Left edge (top-left X) in points. y_pt: Top edge (top-left Y) in points. width_pt: Desired visual width in points (optional; preserves current if omitted). height_pt: Desired visual height in points (optional; preserves current if omitted).

Apply a raw AffineTransform to an element. (1 API call.)

Args: presentation_id: The presentation ID. element_id: The page element ID. transform: An AffineTransform dict (scaleX, scaleY, shearX, shearY, translateX, translateY, unit). unit may be EMU or PT. mode: ABSOLUTE (replace) or RELATIVE (multiply with existing).

Change the front/back stacking order of elements. (1 API call.)

Group two or more elements into a single group. (1 API call.)

Ungroup one or more groups back into individual elements. (1 API call.)

Replace all occurrences of each placeholder string. (1 API call.)

The primary content-fill tool for the template workflow: map showcase placeholder text to real content. All replacements run in one batch.

Args: presentation_id: The presentation ID. mappings: {find: replace} pairs. page_ids: Restrict to these pages (optional; default = whole deck). match_case: Case-sensitive matching.

Insert text into a shape or table cell at a character index. (1 API call.)

Replace ALL of a shape's (or table cell's) text in one call. (1 API call.)

Clears the element's existing text and sets it to text in a single batch (deleteText + insertText). Safe to call on an empty element. This is the ergonomic way to fill a freshly duplicated slide when the showcase example uses real text rather than {{placeholder}} tokens.

Args: presentation_id: The presentation ID. element_id: The shape or table-cell object ID. text: The new full text content.

Clone an entire deck (Drive files.copy), preserving full styling. (1 call.)

The starting point of the high-fidelity template workflow: copies the showcase deck with all masters, layouts, theme and color scheme intact, so the copy can be edited exactly as if you were editing the showcase itself.

Args: source_id: The presentation ID of the template/showcase to copy. title: Title for the new copy. parent_folder_id: Optional Drive folder to place the copy in.

Summarize each slide (layout, element makeup, placeholders, title). (1 call.)

Use after copy_presentation to choose which showcase example slide to reuse for each section of the new deck.

Duplicate a slide within a deck with full fidelity. (1 API call.)

Copies all of the slide's elements and inherited styling. Use it to instantiate a showcase example slide for a new section, then fill it with replace_all_text.

Args: presentation_id: The presentation ID. page_id: The slide to duplicate. insertion_index: Where to place the duplicate (optional).

Delete slides or page elements by ID. (1 API call.)

Use to prune the original showcase example slides after duplicating the ones you want.

Move a block of slides to a new position. (1 API call.)

Hide slides (mark skipped) so they stay as a clone source. (1 API call.)

Keeps the showcase example/original slides in the deck as a reusable "palette" while hiding them from presentation mode. Duplicate from them across as many turns as you like, then prune_parked_slides at the very end.

Unhide slides previously parked (mark not skipped). (1 API call.)

Delete every parked (hidden/skipped) slide — final cleanup. (<=2 API calls.)

Removes ALL slides currently marked skipped, so use it deliberately as the last step once the deck is assembled.

Render a slide to a PNG image for visual verification. (1 expensive call.)

Returns the server-side rendered pixels of the slide — the same image a viewer sees — so you can confirm output looks correct. Rendering is per-page by design (the thumbnail endpoint is an expensive quota operation).

Args: presentation_id: The presentation ID. page_id: The slide object ID. size: LARGE (~1600px), MEDIUM (~800px), or SMALL (~200px) wide.

Render two slides and report their pixel difference. (2 expensive calls.)

Renders both pages, computes a normalized mismatch ratio in [0, 1] (0 = pixel identical), and returns a visual diff image highlighting changed regions. Useful to verify a duplicated/edited slide matches the showcase original.

Returns a text summary followed by the diff PNG.