# Google Slides MCP for presentation workflows. AI Agent Connect > Google Slides MCP gives your AI agent direct access to the Google Slides API. You can instruct your agent to build new presentations from scratch, add specific slide layouts, pull text from existing decks, or perform bulk find-and-replace edits across an entire file. It handles the heavy lifting of deck construction so you can focus on the content. ## Overview - **Category:** productivity - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_ux9PsFUaq8yCClPHvk6hpldJGdfdY4DgZDPFrGe6/ai-agent-connect - **Tags:** google-slides, presentations, slides, deck-building, google-workspace ## Description You can now give your AI client the ability to actually build and manipulate presentations. Instead of copying and pasting text between a chat window and a browser, you just tell your agent what you need. It can open a deck to read the current structure or just grab the plain text to summarize a meeting. If you need to start from zero, the MCP creates a new presentation and populates it with slides using layouts like title, section-header, or title-and-body. Once the slides exist, your agent can fill them with content or fix typos and branding issues across the whole deck in one go. It works through the official Google Slides API, meaning your agent stays within the permissions of the Google account you connect. This turns your AI from a simple writer into a functional presentation assistant that can handle the tedious parts of deck building. ## Tools ### add_slide Adds a new slide to your presentation using layouts like title, section-header, title-and-body, or blank. You can use the returned object ID to fill the new slide with text immediately after creating it. ### create_presentation Starts a brand new Google Slides deck with a title you specify. This is the starting point for turning agent-generated content into a full presentation. ### get_presentation Reads the full structure of a deck, including titles, slide IDs, and object IDs. Use this when your agent needs to understand how a deck is organized before making changes. ### get_presentation_text Extracts only the plain text from a presentation. This is the fastest way for your agent to read, quote, or summarize content without worrying about shapes or layout data. ### replace_slide_text Performs a find-and-replace operation across every slide in the deck. You can use this to update brand names, dates, or titles using plain text or regex patterns. ### set_slide_text Writes text to a specific slide using its object ID. This tool appends text to the end of the existing content on that slide. ## Prompt Examples **Prompt:** ``` Turn these meeting notes into a 5-slide deck. ``` **Response:** ``` Created 'Board sync — Sep 2026' with 5 slides: a title slide, then section headers 'Decisions', 'Open items', 'Risks', and a closing 'Next steps'. Filled each slide with the notes, grouped by theme. ``` **Prompt:** ``` Read my 'Investor update' deck and summarize what each slide says. ``` **Response:** ``` The deck has 7 slides: 1) cover with the quarter, 2) KPIs, 3) product roadmap, 4) enterprise pipeline, 5) new hires, 6) risks, and 7) the final ask. ``` **Prompt:** ``` Replace every 'ACME Inc' in the deck with 'Acme Corp' and make it official. ``` **Response:** ``` Replaced 6 occurrences of 'ACME Inc' with 'Acme Corp' across all slides in one pass. The deck now reads consistently with the new name. ``` ## Capabilities ### Deck Construction Your agent creates new presentations and adds slides with specific layouts. ### Content Extraction The agent reads slide structures or plain text to summarize or analyze existing decks. ### Bulk Text Editing You can run find-and-replace commands to update specific terms across an entire presentation. ### Slide Population The agent inserts text into specific slides using their unique object IDs. ## Use Cases ### Meeting Note Conversion Directly turn a transcript or a list of bullet points into a structured multi-slide deck. ### Brand Consistency Use find-and-replace to update a company name or date across dozens of slides at once. ### Template Automation Generate new decks with pre-set layouts and titles based on specific project requirements. ### Content Auditing Ask your agent to read a deck and summarize the key takeaways or identify risks mentioned in the text. ## Benefits - Eliminates manual copy-pasting between your AI client and Google Slides. - Allows for bulk updates to text and branding across entire decks. - Enables automated deck generation from unstructured notes or data. - Provides a way to programmatically read and summarize existing presentations. ## How It Works Setting up the connection is a one-time process that gives your agent permission to act on your Google account. 1. Create a Google Cloud project and enable the Slides API. 2. Generate OAuth 2.0 client credentials for a web application. 3. Enter your Client ID and Client Secret into the Vinkius interface. 4. Authenticate your Google account through the standard OAuth flow. 5. Connect your AI client to the Vinkius-hosted MCP to start sending commands. ## Frequently Asked Questions **How does the agent access my slides?** The agent uses the credentials you provide to act through the official Google Slides API. It can only access the specific decks that your authenticated Google account has permission to view or edit. **Can the AI change existing text on a slide?** Yes. You can use the replace_slide_text tool to perform find-and-replace across the whole deck, or use set_slide_text to add content to a specific slide. **What kind of slide layouts can be created?** The MCP supports title, section-header, title-and-body, and blank layouts. **Do I need to host the MCP myself?** No. Vinkius hosts and manages the MCP for you. You just provide the credentials and connect your client. **Can my agent read the content of a presentation?** Yes. You can use get_presentation to see the structure and object IDs, or get_presentation_text to pull all the plain text for summarization. **Which Google Cloud permissions are required?** The OAuth flow requests only the Google Slides API scope. The agent can therefore read and modify presentations that your connected Google account can access, and nothing beyond that scope. **How do I get the presentation ID and slide object IDs?** The presentation ID is the long code in the /edit/ URL. Slide object IDs are not visible in the browser — call get_presentation to list them, or add_slide returns the new slide's ID. set_slide_text needs that object ID, not the browser slide number. **Can it rebuild an entire deck from scratch?** Yes: create the presentation, then loop add_slide (choosing a layout each time) and set_slide_text on each returned object ID. This builds a working deck slide by slide. Complex layouts, charts and images are out of scope — this is a text-first workflow.