Creating AI-powered presentations from YouTube videos is an effective way to quickly transform lengthy training recordings, technical panel sessions, or conference talks into structured slide decks. However, when using this workflow, it is essential to clearly understand how the chosen platform operates, whether it accepts direct URLs, and which layer of the video it actually analyzes.
Among official generative AI tools, the workflow splits into two primary approaches: notebook-based models that accept a YouTube link directly as a source, and enterprise presentation tools that generate slides by referencing a cleaned text transcript stored in Google Drive.
Can a YouTube link be turned directly into a presentation
Directly processing YouTube links with AI tools is possible, but it comes with distinct technical boundaries. Under the Gemini Notebook documentation on adding sources, only URLs of public YouTube videos that have user-uploaded or automatically generated captions can be imported as sources. Videos without speech or without captions are not supported.
By contrast, in the workflow for creating presentations with Gemini in Google Slides, instead of pasting an external YouTube link directly into the prompt box, the primary approach relies on linking reference files through Google Drive or augmenting the prompt with web search. Consequently, the tool you choose directly dictates your workflow when working with video content.
If your goal is not to summarize the video's content into a newly designed slide deck, but rather to insert a playable video element into an existing presentation, you can follow the embedding steps in our guides on adding a YouTube video to PowerPoint or embedding a YouTube video in Google Slides.
The difference between speech transcripts and video frame analysis
The most common misconception users encounter is assuming that when an AI receives a video URL, it watches and analyzes every visual frame, on-screen slide, or graphic illustration individually.
In official systems, adding a YouTube URL imports only the video's text transcript into the platform. Visual frames, animations, or diagrams appearing momentarily on screen are not processed directly. Furthermore, freshly uploaded videos less than 72 hours old may not yet be processable, and the video's caption file must contain fewer than 500000 words. If the video is deleted or set to private, the imported source is removed from your workspace within 30 days.
Caption availability and the accuracy of automated transcription vary from video to video. For this reason, if you are working with a video centered on on-screen code blocks or drawings, you must account for the fact that only the spoken words will be reflected on the slides.
Generating a slide deck from a YouTube URL with Gemini Notebook
To turn a public YouTube video link directly into a presentation, you can follow the steps for creating a Slide Deck with Gemini Notebook:
- Sign in to Gemini Notebook on your computer or device and create a new notebook.
- Click the Add sources option on the left side and enter the URL of a public YouTube video with captions.
- Once the source is successfully added, select the Slide Deck option located in the Studio panel on the right.
- Configure the format, output language, slide length, and the source to use. Write a guiding prompt describing your desired presentation style.
- Click the Generate now button. It may take Gemini Notebook a few minutes to generate the slide deck; you can continue working in the notebook in the meantime.
The generated slide deck can be presented in full screen directly within the platform, downloaded as a PDF, or exported as a PowerPoint (.pptx) file. However, keep in mind that this slide deck generation feature is available for users aged 18 and older, and AI outputs may contain factual inaccuracies.
Controlled generation using manual transcripts and Google Slides
When you want complete editorial control over the source text instead of adding a YouTube link directly, extracting the transcript through the YouTube interface is a reliable alternative:
- On the YouTube watch page, expand the description box and click Show transcript to open the transcript panel.
- Copy the spoken text displayed in the panel.
- Paste the text into a Google Doc and clean up irrelevant sections such as sponsor announcements, greetings, or calls to subscribe.
- Open a new blank presentation in Google Slides and click the Ask Gemini icon in the top right.
- Select your cleaned document via Add sources or Add from Drive, or reference your file in the prompt box using the @ symbol.
- Review the presentation plan generated by Gemini (overview, sources, steps), edit the title and steps if necessary, and click Approve to generate the presentation. Do not close the tab until generation is complete.
Plan requirements and language support restrictions
The feature to generate full presentation decks from scratch with Gemini in Google Slides requires an eligible Google Workspace or Google AI subscription. On personal accounts, it is offered through the Google Workspace Experiments program.
As an important technical constraint, creating an end-to-end new presentation deck via the Gemini side panel in Google Slides is currently supported only on desktop computers and exclusively in English. Therefore, when planning a full deck in the interface, prompts must be submitted in English; Turkish prompts are not yet officially supported in this full-deck generation workflow. For single-slide text and image generation or rewriting tasks, other supported Workspace languages can be used.
Example scenario: Generating a presentation from a technical video transcript
In a hypothetical scenario, consider a 12-minute technical talk on YouTube titled "Database Optimization in Web Applications." After cleaning the transcript text and uploading it to Drive, a prompt structured like this can be entered into the English-language Slides Gemini panel:
Prompt Example: Create a 4-slide presentation based on the referenced file @Database_Optimization_Transcript for a technical team. Focus strictly on system bottlenecks, indexing strategies, caching layers, and connection pooling. Do not include assumptions not present in the transcript.
Generated Outline Flow: - Slide 1: Introduction and System Bottlenecks (Latency and key metrics emphasized in the transcript) - Slide 2: Indexing Strategies (The role of B-tree indexes and the cost of redundant indexes) - Slide 3: Caching Architecture (In-memory caching tiers that alleviate database load) - Slide 4: Connection Pool Management (Preventing connection exhaustion and concluding recommendations)
Encountered limitations and common mistakes
- Overlooking automatic caption errors: Technical terminology, personal names, or acronyms in YouTube's automated captions may contain transcription errors. These inaccuracies can carry over directly into slide titles.
- Disregarding video privacy status: Gemini Notebook supports only public videos. Unlisted or private videos cannot be added via link.
- Closing the tab prematurely: Once presentation generation begins in Google Slides, you can switch to other tabs or windows, but you must not close the Google Slides tab or the Gemini side panel.
- Video playback issues: If embedded video clips inside slides fail to play after the presentation is generated, check permissions or link settings; in such cases, consult the troubleshooting steps in our guide on PowerPoint YouTube video not playing.
Verification checklist
- Did you verify that the video is public and has an available caption track?
- If using Gemini Notebook, did you select Slide Deck output from the Studio panel?
- If using Google Slides, did you confirm your account has an eligible Workspace or Google AI plan?
- When generating a full deck in Slides, did you account for the desktop and English-language requirements?
- Did you clean sponsor mentions and greetings from the transcript?
- Did you keep the working browser tab open until generation finished?
Frequently asked questions
Can AI read all the visual slides in a YouTube video directly?
No. The official Gemini Notebook integration pulls only the spoken transcript of the YouTube video as a source. Graphics, whiteboard drawings, or presenter slides in the video frames are not included in this import.
Can the generated slides be converted back into a video using AI later?
Yes. Through the capability for converting Google Slides into Google Vids, you can import your existing presentation into Google Vids, convert slides into scenes, and turn them into a video complete with AI voiceovers and background music.
← Back to blog
