Getting started with Auto-Snapper-AI

What is Auto-Snapper-AI?

Auto-Snapper-AI is a desktop app for macOS and Windows that turns any window on your screen into a personal tutor. You pick a window from your own study materials — a PDF, a slide deck, a practice-question set, a web page — and the app automatically takes a snapshot every few seconds, sends it to the AI of your choice, and shows you a clear explanation. You learn without ever switching tabs to ask a chatbot.

It's ideal for

• Certification and exam prep — drill concepts and work through practice questions from materials you own.
• Self-paced learning — get plain-language explanations of dense documentation or coursework.
• Review sessions — capture a whole study session and export it for a second pass.

Quick start (5 minutes)

1. Install and open the app, then complete the short setup wizard.
2. Activate your license — accept the terms and paste the key from your purchase email.
3. Add your AI key — in Settings, paste your OpenAI and/or Anthropic API key and save.
4. Test the connection — click Test API Connection to confirm everything works.
5. Grant screen permission (macOS) — click Select Window… and approve the Screen Recording prompt.
6. Pick your study window, choose a persona, and click Start Continuous.

That's it — answers start flowing as you study.

Setting up

Activate your license

After purchase you'll receive a license key by email. On first launch, accept the terms and paste the key. It works fully offline for your subscription term, so once you're activated you can study with no internet connection. When your term ends, paste a renewal key on the 'license expired' screen — the app reminds you as the date approaches.

Add your AI key

Auto-Snapper-AI uses your own OpenAI or Anthropic API key, so you pay your AI provider directly and choose your own model. In Settings, paste your key and click Save. Keys are stored in your operating system's secure keychain — never in plain files and never sent to us.

Test the connection

Click Test API Connection. The app sends a tiny test request that exercises exactly what a real capture would — so if a key is wrong or a model setting is off, you find out now instead of mid-study.

Grant screen permission (macOS)

The first time you click Select Window…, macOS asks for Screen Recording permission. Approve it (System Settings → Privacy & Security → Screen Recording), then relaunch the app. Without this, the app can't see your study window. On Windows there's no extra permission step.

Your first study session

Pick a window

Click Select Window… and choose the window with your study material. A live thumbnail appears so you can confirm you picked the right one. The app captures only that window — nothing else on your screen.

Choose a study persona

A persona is a ready-made tutor profile: a tailored prompt plus its own reference material. Pick one from the Persona dropdown to set the tone and subject. The app ships with personas for several technical certification tracks, plus a generic all-purpose tutor — and you can create your own.

Start studying

Click Start Continuous. The app captures your window on a timer (default every 30 seconds) and shows each explanation in the Latest Answer panel. Keep your eyes on your materials — the app does the asking for you.

• Capture Now — grab one snapshot immediately, any time.
• Pause / Resume — take a break without ending the session.
• Stop — end the session, or press the global hotkey (default Cmd/Ctrl + Shift + S) from anywhere.

Making answers smarter

Feed it your own notes (Reference Context)

This is the feature that makes answers match your course. Point a persona at a folder of your study materials — notes, study guides, documentation — in .md, .txt, .json, or .xml format. The app reads them and treats them as authoritative on every capture, so the AI answers from your syllabus instead of guessing.

1. Browse to your folder of notes — the app ingests them automatically.
2. Edit your notes anytime, then click Re-ingest to refresh.
3. PDFs aren't supported — convert them to text or Markdown first.

Read the confidence signal

• High — a clean, bold answer. The AI is sure.
• Medium — a 'review this' note with the AI's reasoning shown.
• Low — a clear warning that the AI wasn't sure. Double-check these.

Get a second opinion (Voting)

Turn on Voting to have both OpenAI and Anthropic answer each capture. If they agree, you can be confident. If they disagree, the app flags it — a great signal that the topic is genuinely tricky and worth more study. (Voting uses both providers, so it costs more per capture; use it for high-stakes review.)

Let it retry the hard ones

With auto-retry on low confidence enabled, the app automatically takes another pass at shaky answers, using your recent confident answers as extra context. Sensible caps keep costs in check.

Reviewing and exporting

Session history

Every capture and answer is saved locally during your session. Scroll back, click any entry to re-read the full explanation, and search across the whole session — all stored privately on your own machine.

Export your study notes

• Study bundle — a self-contained folder with the captured images, per-question notes, and a consolidated summary. Best for a thorough second pass.
• Markdown / JSON / CSV — single-file exports for quick sharing or importing into other tools. CSV opens cleanly in a spreadsheet for tracking weak spots.

Personalizing with your own personas

Want a tutor for a subject the built-ins don't cover? Create your own:

1. Go to Settings → Personas → + New persona.
2. Name it (e.g. 'AWS Solutions Architect'), optionally clone an existing persona's prompts, and click Create.
3. The persona's reference folder opens — drop in your study notes and docs.
4. Click Re-ingest, then select your new persona from the dropdown.

You can also add materials to the built-in personas the same way — open the persona's folder, drop files in alongside the bundled docs, and re-ingest. Your edits and extra files are always kept; updates never overwrite them.

Privacy and your data

• No middle-man. The app talks only to your chosen AI provider. We run no server between you and the AI.
• Keys stay secure. Your API keys live in your OS keychain, never in plain files or logs.
• Your data stays local. Captures and history are stored on your machine. One button in Settings clears all local history.
• Pixels only. The app captures images of the window you pick — it doesn't scrape browsers or read other apps' data.

Tips for better study sessions

• Match the interval to the task — slow it down for deep reading; speed it up for rapid review.
• Load your notes first — Reference Context is the single biggest lever on answer quality.
• Watch the confidence colors — spend your time on the medium and low ones.
• Use the live preview to make sure you've selected the right window before you start.
• Pick the right model — a fast, cheap model for review; a stronger reasoning model for difficult material.
• Export at the end so your session becomes a reusable study resource.

Troubleshooting

The window list is empty (macOS)

Screen Recording permission isn't granted. Open System Settings → Privacy & Security → Screen Recording, enable Auto-Snapper-AI, and relaunch.

A test or capture says my key is invalid

Re-check the key in Settings (no extra spaces), make sure the matching provider is selected, and that your AI account has available credit. Then click Test API Connection again.

It keeps skipping captures

Turn off duplicate-skipping in Settings if you're studying screens that look similar but contain different questions.

An answer looks wrong

Add the relevant material as Reference Context so the AI answers from your sources, and check the confidence rating — low-confidence answers are flagged for a reason.

A final reminder

Auto-Snapper-AI is here to help you learn and prepare. Use it on your own materials, to study for your exams and certifications — not during live, graded, or proctored assessments. Used the right way, it's like having a patient tutor read over your shoulder. Happy studying.