About ScribeShot

ScribeShot is a local-first macOS app for turning YouTube videos into structured notes, screenshots, chapters and mind maps — then searching, asking, and testing yourself across the knowledge you have saved. It was built by Disha Kalra.

Why this exists

There are plenty of tools that will give you a YouTube summary. Most of them return the same thing: a few bullet points extracted from the transcript, the spoken words without any of what was on screen. For a lecture where the diagram on the whiteboard is the whole point, or a coding tutorial where the useful part is in the terminal window, that is not actually useful.

The specific problem ScribeShot is built to solve is the gap between watching and having something to keep. Watching is slow and passive. A summary is too thin and has no timestamps. What works is structured notes with the visuals captured in context — the screenshot filed under the heading it belongs to, not in a separate folder of anonymous images.

Once you have more than a handful of useful video notes, a second problem appears: remembering which video contained the idea you need. That is why ScribeShot now treats the whole saved library as a personal knowledge base. You can search across videos and matching chapters, ask questions grounded in your saved material, and use Test Me for active recall without reopening every source one by one.

The second thing that bothered me was the subscription model. Learning tools charge monthly even though a student's intensive use is seasonal and a developer's use is sporadic. If a tool is genuinely useful it should cost once, not indefinitely, and the files it produces should keep working without the app installed.

The design principles

• Local-first. Notes, screenshots, transcripts, and chat history stay on your machine. Nothing is uploaded to a ScribeShot server. The app works offline once the model is downloaded.
• Plain files. Output is standard Markdown and image files in a folder you choose. They are readable in any editor, importable into any vault, and yours permanently.
• Knowledge should compound. Saving the twentieth video should make the first nineteen more useful. Library-wide Search, Ask, and Test Me turn individual notes into something you can retrieve and learn from as a collection.
• One-time price. A lifetime licence means no recurring cost and no incentive to make the format proprietary.
• macOS quality. ScribeShot is a Mac-first desktop application designed around focused video learning. The screenshot workflow, local-model integration, video player, and local-file library are built to feel at home on macOS.
• Honest about limitations. YouTube only. macOS only for now. Transcripts must exist. Videos are still added individually even though retrieval works across the saved library. These are real constraints and they are documented rather than hidden.

Who it is built for

The people who get the most out of ScribeShot tend to have a specific problem in common: they watch long, information-dense videos and need something to show for it. That includes students working through recorded lectures, developers who learn from conference talks and screencasts, researchers keeping notes on long-form interviews, and anyone who finds rewatching a video to find one thing an unacceptable way to spend time.

The useful unit is no longer only one processed video. A semester of lectures can become something you search and test before an exam; a folder of conference talks can answer the question you half remember from months ago; and a growing tutorial library can stay useful without depending on memory of filenames or titles.

The ADHD and focus-related use case was not something I originally planned for, but it is one of the most consistent pieces of feedback: chapters and AI Video Highlights let someone work through a long video in short passes, picking up where they left off without losing the thread.

Get in touch

The best place to ask questions, report bugs, or suggest something is the ScribeShot Discord community. I read everything there.