Make any video understandable by AI

Right now, you can't just hand a video to ChatGPT, Claude, or Gemini and ask "what happened in this?". We change that. Paste a link from YouTube, Instagram, TikTok, Facebook, X, or LinkedIn (or upload a file) and you'll get back a single PDF those AI tools can actually read end to end, including both the words spoken and what was on screen.

Picking up where you left off…
Need to save a class recording first? See the FAQ →
Advanced options
1 / 5s
Lower = denser frames (good for tutorials). Higher = sparser (good for talking heads).
Higher = clearer code/slides. Larger PDF size.
We'll size the PDF to your AI's upload limit — adjusting screenshot density on long videos if needed.
Include in output
YouTube
Instagram
TikTok
Facebook
X
LinkedIn
Vimeo
Reddit
Twitch
Canvas (Click Here for Details) + Direct Upload
How it works

Three steps. One file out.

No setup, no API key, no credit card for the free tier (30 min / month). Drop in a video, get a PDF that any AI can read.

1

Paste or upload

A link from YouTube, Instagram, TikTok, Facebook, X, LinkedIn, or drop a file straight in.

2

We process

We download the video, transcribe the audio with speaker diarization, and capture timestamped screenshots at scene changes.

3

Download the PDF

One file, ~1 - 5 MB. Drag it into Claude, ChatGPT, or Gemini. The AI reads both the transcript and the visuals.

Why it works

Built so AI actually gets it

Most "video to PDF" tools dump screenshots at random intervals and call it a day. We're built for one thing: making sure AI tools like ChatGPT, Claude, and Gemini can understand both what was said and what was on screen.

A

Word-perfect transcripts

Top-tier speech recognition catches every word with proper punctuation and labels who said what.

★

Smart screenshots

We grab a frame whenever the screen actually changes (a new slide, a new code window, a different scene), not at random timestamps. Near-duplicate frames get filtered out automatically.

↔

Words and visuals in sync

Every screenshot lands in the transcript at the exact moment it appeared on screen. The AI reads each visual right where it belongs.

⌘

Fits any AI's upload limit

Tuned to stay under each chat tool's PDF size cap (Claude 32 MB, ChatGPT ~30 MB, Gemini 2 GB). Every output drops into one upload.

⊕

Tune it for your video

Sliders for screenshot density and image quality. More frames for tutorials with rapid changes, fewer for talking-head content.

∅

Your content stays private

Source videos and outputs auto-delete after download. Nothing stored, nothing trained on.

FAQ

Common questions

What platforms does the link option support?
YouTube, Instagram, TikTok, Facebook, X (Twitter), LinkedIn, Vimeo, Reddit, Twitch, Bluesky, Streamable, and ~20 others. Public videos only. Anything behind a login or paywall won't work via URL. For those, just download the file yourself and use the upload option.
I have a video file already. What's the easiest way to upload it?
Drag it onto the upload tab on the homepage, or click the dropzone to pick a file. We accept MP4, MOV, MKV, WebM, AVI up to ~5 GB. The conversion runs in the background, and you'll get a download link when it's done.
I have authorized access to a video on Canvas, Panopto, Kaltura, or another class-recording site. How do I save it as a file to upload here?
Most class-recording players don't expose a download button, even when the institution allows you to save the content. Several free, third-party browser extensions let you capture authorized video streams to a local MP4 file:
  • Video DownloadHelper: available for Chrome and Firefox. This is the most reliable option if you're not on Chrome.
  • FetchV: Chrome only. FetchV sometimes saves audio and video as separate files. Upload them both and we will combine them ourselves.
  • Stream Recorder: Chrome only.

Safari: none of these extensions have Safari versions. If you're on a Mac, install Chrome or Firefox for this step.

Double-check before you download. Review your institution's acceptable-use policy and the video host's terms of service before saving a recording. You're responsible for ensuring you have the right to save and process the content you upload here.

ProForma Concepts is not affiliated with these tools and does not develop or distribute them. Use of any third-party extension is governed by its own terms.

Why a PDF and not a Word doc or markdown file?
PDF is the only single-file format that every major AI chat tool (ChatGPT, Claude, Gemini) can open and read both the text and the images from. Word docs lose the visuals. Markdown with embedded images doesn't work either, because chat apps treat .md files as plain text and ignore the image data. PDF is what works.
How long can the video be?
Practical ceiling is around 2 hours. We dynamically tune output size to stay under Claude's 32 MB PDF limit. For longer content, the upload path always works, with no hard limit on processing time.
What's the transcription quality like?
We use ElevenLabs Scribe, one of the top speech-recognition systems available today. It catches every word with proper punctuation and labels who said what (so you can tell speakers apart in interviews and conversations).
Is my content stored?
Source videos are deleted from our server within 24 hours (and immediately after processing in most cases). Output PDFs and ZIPs auto-delete after you download them, with a 24-hour S3 lifecycle as a safety net. We don't index or train on your content.
Is this free?
Yes, the free tier includes 30 minutes of video per month with no credit card. Perfect for one-time use. Paid plans are billed by total minutes, not video count: Starter $10/mo (600 min), Pro $30/mo (3,600 min, ~30 lectures), and Unlimited $60/mo (unlimited minutes, a full med-school lecture load). Cancel anytime from your account.