ScribeShot
YouTube source access
ScribeShot needs network access to open the source, fetch available transcript and metadata, and play the YouTube video.
Offline YouTube Summarizer for Mac
Fetch a YouTube video and its available transcript, then run the supported AI work locally on your Mac: summaries, chapters, Markdown notes, screenshot understanding, mind maps and video chat. The source still needs the internet; the AI processing does not need a cloud AI provider.
Local-first, with a clear network boundary
What needs the internet and what can stay on your Mac
YouTubeScribeShot needs access to the source video, metadata, and available transcript.
Source fetchedThe transcript and captured frames become inputs to the ScribeShot workspace.
On your MacGemma 4 can generate notes and chapters, analyze screenshots, and answer video questions locally.
Model setup: download a supported local model before generating notes. Supported AI processing runs on your Mac without an API key.
What stays local, and what still needs the network
AI processing
After a supported model is downloaded, supported AI inference runs locally on your Mac.
Your output
Notes, transcripts, screenshots, chats and learning artifacts remain ordinary files inside your chosen ScribeShot library.
Choose the local model that fits your Mac
| Model | Approx. download | Best fit |
|---|---|---|
| Gemma 4 E2B | ~2.0 GB including vision support | Lightweight local processing; a good starting point for an 8 GB Mac. |
| Gemma 4 E4B (default) | ~3.6 GB including vision support | Best general balance for everyday notes, chapters and visual understanding. |
| Qwen3-VL 8B | ~5.8 GB including vision support | Higher-quality local vision, OCR and reasoning; 16 GB+ recommended. |
What “offline” does and does not mean
The useful distinction is between source access and AI processing. YouTube itself is an online source, so ScribeShot cannot fetch a new video or transcript with the network disconnected. Once the source material and model are present, supported AI generation and analysis can run locally without an AI API key.
Why use local AI for video notes?
- No AI API key: there is no cloud-AI account or key to configure.
- Predictable cost: local inference has no per-request AI charge.
- Local processing boundary: supported AI work runs on your Mac.
- Model choice: pick a lighter Gemma model for efficiency or Qwen3-VL when higher-quality visual reasoning matters more.
Common questions
Do my transcripts or notes get sent to cloud servers?
No. In local mode, all processing runs on your Mac GPU/CPU and data stays on your disk.
What Mac hardware do I need?
An Apple Silicon Mac (M1/M2/M3/M4) with 8GB RAM for Gemma 4 E2B or 16GB+ RAM for Gemma 4 E4B.
Is there a monthly subscription for local AI?
No. ScribeShot is a $29 one-time lifetime purchase.
Process video notes with local AI
Fetch the YouTube source, then keep the AI processing on your Mac with a supported Gemma 4 model.
Try 3 videos free • 3 screenshots per video • No credit card • Zero AI cost with local LLM