Video

Convert Video to Slides: Extract Key Frames into Presentations (2026)

S
Sharayeh Team
β€’
10 min read
β€’

🎬 Try it now:
Video to Slides β†— | YouTube to Slides β†—


Convert Video to Slides

Recorded lectures, webinars, training sessions, and conference talks contain valuable visual content trapped inside video files. Extracting the slides β€” along with the presenter's spoken content β€” gives you a reusable, editable presentation that's searchable, printable, and easy to share.


How AI Extracts Slides from Video

The AI uses multiple techniques:

1. Scene Detection

The AI analyzes frame-by-frame changes to detect when a new slide appears. It captures:

  • Slide transitions β€” when the visual content changes significantly
  • Partial updates β€” when bullet points are added incrementally
  • Screen shares β€” presentation slides shown during screen recordings

2. OCR (Optical Character Recognition)

Text visible in the video frames is extracted using OCR:

  • Slide titles and body text
  • Code snippets (for tech talks)
  • Diagram labels
  • On-screen annotations

3. Speech-to-Text

The audio track is transcribed and:

  • Aligned with corresponding slides
  • Added as speaker notes in the PowerPoint output
  • Used to generate slide summaries

Step-by-Step Guide

Method 1: Direct Video Upload

  1. Go to Video to Slides
  2. Upload your video (.mp4, .webm, .mov, .avi)
  3. The AI processes the video:
    • Detects slide changes β†’ captures key frames
    • Runs OCR β†’ extracts text from each frame
    • Transcribes audio β†’ creates speaker notes
  4. Review the generated slides
  5. Download as .pptx

Method 2: YouTube URL

  1. Go to YouTube to Slides
  2. Paste the YouTube video URL
  3. The AI downloads and processes the video automatically
  4. Same output: slides with extracted text and speaker notes

Processing Time

Video Length Approximate Processing
5 minutes 30 seconds
15 minutes 1–2 minutes
30 minutes 2–4 minutes
60 minutes 5–8 minutes
2+ hours 10–15 minutes

Supported Video Formats

Format Max File Size
MP4 (H.264) 2 GB
WebM 2 GB
MOV 2 GB
AVI 2 GB
MKV 2 GB

Best Use Cases

Lecture Recordings

  • Extract all slides from a recorded lecture
  • Get the professor's explanations as speaker notes
  • Create a study guide from the presentation

Webinar Recordings

  • Convert marketing webinars into shareable slide decks
  • Extract key insights for social media posts
  • Create training materials from recorded sessions

Conference Talks

  • Capture slides from recorded conference presentations
  • Build a library of presentation content from talks you've attended
  • Share insights with team members who couldn't attend

Training Videos

  • Convert training recordings into reference materials
  • Create printable handouts from video content
  • Build a knowledge base from recorded sessions

Screen Recordings

  • Extract the presentation from a Zoom/Teams recording
  • Separate the slide content from the video call interface
  • Clean up meeting recordings into shareable decks

Optimizing Video Quality for Extraction

Recording Tips

For future recordings, optimize for slide extraction:

  1. Record at 1080p minimum β€” Higher resolution = better OCR
  2. Use high contrast slides β€” Dark text on light background
  3. Minimize presenter overlay β€” Full-screen slide view when possible
  4. Stable frame rate β€” 30fps is sufficient
  5. Clear audio β€” Good microphone for better transcription
  6. Pause between slides β€” 2-second pause helps detection

Handling Poor Quality Videos

For low-resolution or noisy videos:

  • The AI uses enhanced OCR with noise reduction
  • Speaker notes may have lower accuracy
  • Consider uploading the original slides if available and using the video just for speaker notes

Advanced Features

Slide Deduplication

When presenters go back and forth between slides, the AI:

  • Detects duplicate frames
  • Keeps only unique slides
  • Orders them logically

Incremental Slide Building

For slides that build up content (animations):

  • The AI captures the final state of each slide
  • Or optionally captures each step as a separate slide
  • You choose which approach in the settings

Multi-Speaker Detection

For panel discussions or multi-presenter videos:

  • The AI identifies different speakers
  • Tags speaker notes with speaker attribution
  • Can create section breaks between speakers

Frequently Asked Questions

Can it extract slides from Zoom recordings?

Yes. Zoom's MP4 recordings (including gallery view and speaker view) work well. The AI focuses on the shared screen content and extracts slides from it.

What about videos without visible slides?

For talking-head videos without slides, the AI creates slides from the speech content β€” summarizing key points into visual slides. It works similarly to the Text to Slides tool.

Does it work with videos in languages other than English?

Yes. OCR and speech recognition support Arabic, French, Spanish, German, Turkish, Japanese, and more.

Can I extract only specific parts of a video?

Currently, the full video is processed. After extraction, you can delete unwanted slides from the generated deck.


Related Tools

Share this article: