cancel_generation_run
ChatGPTStop a run started by generate_project. Pages already written stay in the deck. Returns the run's status; poll get_generation_run until its phase is failed (the run ends canceled). A run that has already finished cannot be canceled.
create_theme
ChatGPTCreate a custom presentation theme in the user's Encelade workspace from brand colours, fonts and styles, e.g. when the user asks to send their brand or design system to Encelade. To change a theme the workspace already has, use update_theme instead. name and colors.accent are required; anything left out is derived from the given colours or kept from Encelade's default theme. Colours must be hex. For the logo pass exactly one of: logo.url, a public link to the logo file the source design uses; logo.svg, its SVG markup when it has no public link; or logo.brandDomain, the brand's website, to look up its published logo. Prefer the source design's own logo over brandDomain. For a font Encelade doesn't offer, pass a direct link to its file as fonts.heading.url or fonts.body.url. pageBackground, contrast.pageBackground and each backgroundVariants fill also take an image: a public https link Encelade downloads. Generated decks cycle through pageBackground and the backgroundVariants. If the user has a brand-guidelines PDF or PPTX, the theme studio can instead create a theme from that file, including its backgrounds, at https://www.encelade.ai/brand/themes/new. The response lists what was derived and anything that could not be applied, so you can tell the user. Returns themeId: pass it as theme to generate_project or plan_project.
delete_project
ChatGPTPermanently delete a deck by its pid. This action cannot be undone. By default also removes uploaded media exclusively owned by you and no longer referenced by any surviving deck; pass deleteAssets: false to keep those assets.
delete_theme
ChatGPTPermanently delete a custom theme from the user's Encelade workspace, for everyone in it. Get its themeId from list_themes. There is no undo. If decks use the theme, the first call deletes nothing and returns confirmationRequired with decksUsingTheme: tell the user the theme's name and how many decks use it, and ask whether to delete it. Only if they agree, call again with confirmDecksUsingTheme set to that count. If the count changed in between, you get confirmationRequired again with the new count; ask the user again. Generated decks keep the look the theme had when they were created; new decks can no longer use it.
edit_project
ChatGPTEdit an existing deck the way the editor's AI chat does: rewrite or translate text, restyle layers, add, reorder or delete pages, change the theme. Name the pages to change in the instruction, or pass pageId for a request about one page. Changes are saved unless dryRun is true. Returns status (applied, proposed, no_changes or truncated), the assistant's message, the changes made and anything skipped. A truncated run saves nothing: split the request. Uses AI credits, dry runs included.
generate_project
ChatGPTGenerate a full presentation from a topic and outline hints. Returns a session ID immediately — poll get_generation_session to track progress. Required fields: outlineHints (array of slide topic strings). Planning typically takes 30–90 seconds; generating takes 60–180 seconds depending on slide count. Do not flag a session as potentially stuck until 5 minutes have elapsed since startedAt. Optional model param selects the LLM — one of: gpt-6-astra, gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna, gpt-5.5, gpt-5.4, gpt-5-mini, claude-sonnet-5 (default), claude-opus-5-5, claude-sonnet-4-6, claude-sonnet-4-5-20250929, claude-opus-4-8, claude-haiku-4-5-20251001, gemini-3.1-pro-preview, gemini-3.8-flash, gemini-3.5-flash-lite, gemini-3.5-flash, gemini-3.1-flash-lite, gemini-2.5-pro, gemini-2.5-flash.
get_generation_run
ChatGPTPoll a run started by generate_project. Returns a short status line (planning / generating / done / failed) with a link to the deck; the structured payload also carries the phase, the run's stage and its plan. A run takes a few minutes: re-poll with the same runId about every 15 seconds until the status line starts with Done. or the phase is failed. Once done, call get_project once with projectPid to open the interactive slide preview. This status tool does not open a preview.
get_generation_session
ChatGPTPoll the status of a presentation generation session. Returns a short human-readable status line — planning / generating / done / failed — plus a link to the deck once it is available. The structured payload also exposes the current phase, events, plan, and link for programmatic clients. Planning typically takes 30–90 seconds; generating takes 60–180 seconds depending on slide count. Re-poll with the same sessionId until the readable status starts with Done. Once phase is done and projectPid is present, call get_project once with that pid to open the interactive slide preview. This status tool does not open a preview.
get_project
ChatGPTGet a single presentation deck by its pid (public ID). Opens an interactive slide preview in hosts that support MCP Apps.
get_slide_preview
ChatGPTRender a private slide preview for the deck viewer.
list_projects
ChatGPTList presentation decks. Returns paginated results. Use the returned cursor to fetch the next page.
list_themes
ChatGPTList the custom presentation themes in the user's Encelade workspace, newest first. Pass a theme's themeId as theme to generate_project or plan_project to build decks in it. Check this before create_theme so you don't recreate a brand theme the workspace already has; to change one, pass its themeId to update_theme.
personalize_template
ChatGPTClone an existing template into a new deck and rewrite its content for a specific customer using your instructions and data (Mode C). Layout is preserved — only text and in-place imagery change. Returns a session ID immediately — poll get_generation_session to track progress. Required field: templateId (a template UUID). Optional: instructions (free-text guidance), projectName, useConnectors (pull from your connected data sources). Personalization typically takes 60–180 seconds depending on deck size; do not flag a session as stuck until 5 minutes have elapsed since startedAt.
plan_project
ChatGPTGenerate an outline/plan for a presentation without creating the full content. Returns a plan that can be reviewed before generation. Planning typically takes 30–90 seconds. Do not flag a session as potentially stuck until 5 minutes have elapsed since startedAt. Optional model param selects the LLM — one of: gpt-6-astra, gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna, gpt-5.5, gpt-5.4, gpt-5-mini, claude-sonnet-5 (default), claude-opus-5-5, claude-sonnet-4-6, claude-sonnet-4-5-20250929, claude-opus-4-8, claude-haiku-4-5-20251001, gemini-3.1-pro-preview, gemini-3.8-flash, gemini-3.5-flash-lite, gemini-3.5-flash, gemini-3.1-flash-lite, gemini-2.5-pro, gemini-2.5-flash.
update_project
ChatGPTUpdate an existing deck. Pass the project pid and the fields to update.
update_theme
ChatGPTChange an existing custom theme in the user's Encelade workspace. Get its themeId from list_themes. Only the fields you pass change; everything else is kept. chartColors and backgroundVariants replace the whole list, and backgroundVariants: [] removes every variant. Pass null to remove titleCard, contrast or logo. A logo without brandDomain, url or svg only moves or resizes the current one. Fields take the same values as in create_theme, including font and image links. The change applies at once to every deck using the theme, and there is no undo: confirm a large change with the user first, and tell them how many decks it restyled (decksUsingTheme). The response lists what changed, what changed as a result (derived), and anything that could not be applied.