You install Vzlyze, capture a screenshot, and hit the analysis panel. There are four AI modes staring back at you: Quick Analysis, Research Mission, Compare Screenshots, Debug UI. They each cost credits. They each promise something a little different. And the naming is intentional but not instantly self-explanatory.
This guide is the shortest path from "I have a screenshot" to "I picked the right mode." Skip to the decision tree if you want the answer in ten seconds, or read the breakdown if you want to understand why each one exists.
The ten-second decision tree
Ask yourself, in this order:
- Are you looking at a UI you're building or auditing? Use Debug UI.
- Do you have two screenshots you want compared side by side? Use Compare Screenshots.
- Do you want deep research with web search, sources, and follow-ups? Use Research Mission.
- None of the above? Use Quick Analysis for a fast summary.
That's the whole thing. Everything below is context on when each answer is right.
Quick Analysis
Quick Analysis reads your screenshot and returns a plaintext summary of what it sees plus a small handful of observations and follow-up questions. No web search, no comparisons, no source lookups. Around ten seconds end to end.
Best for: a page you saw and want a fast take on. An email thread you want summarised. A chart whose caption is missing. A tweet you want context on without leaving the page.
Skip if: you need external context (that's a Research Mission), you're comparing to another screenshot (that's Compare), or you're a Pro or Business user (your plan replaces Quick with the deeper modes and Quick Analysis is disabled).
Research Mission
Research Mission is Quick Analysis with the internet turned on. It reads your screenshot, decides what's worth looking up, runs live web searches, and returns a structured report: what you're looking at, verified facts pulled from the web, comparisons where relevant, recent context, and cited source links. Runs in about thirty seconds because it's making real network calls.
Best for: product research (specs, pricing, reviews from sources you didn't screenshot). Article context (who wrote this, when, is it disputed). Comparisons to a market benchmark. Anything where the answer isn't in the screenshot but the screenshot points at what to look up.
Skip if: the screenshot already contains everything you need to know (Quick is faster and cheaper). You're on Starter (Research Missions require Pro or Business). You already know the topic and want raw summarisation only.
Compare Screenshots
Compare Screenshots takes two captures at once and returns a side-by-side analysis with a real recommendation at the end. It's the only mode that operates on two inputs. You capture the first thing, toggle Compare, capture the second thing, then run.
Best for: product A vs product B. Version 1 of your landing page vs version 2. Your pricing page vs a competitor's. Any question that starts with "which one is better" or "what changed between these."
Skip if: one screenshot is enough to answer your question. You're not on Business (Compare is a Business-only feature; there's no Free trial). You want the analysis to include external web research (that's a Research Mission, not Compare, which stays inside the two images you gave it).
Debug UI
Debug UI is a frontend audit tool. It captures the page, extracts up to forty computed-CSS observations from the live DOM, and returns a five-section report: what's shown, bugs identified, root causes, suggested CSS fixes, and priority. It's the only mode that pulls structured DOM data alongside the image.
Best for: a page you're building where something feels off but you can't put your finger on it. A layout that behaves badly on specific viewport widths. An accessibility check (contrast ratios, focus rings, missing labels). Handing a bug report to a designer or a junior developer.
Skip if: you want narrative analysis of a non-UI image (that's Quick or Research Mission). The bug is runtime or state-dependent (dropdowns that break when opened, console errors, race conditions): those don't show up in a static capture. See the Debug Screenshot deep-dive for its full envelope.
Cost cheat sheet
| Mode | Free | Starter | Pro | Business |
|---|---|---|---|---|
| Quick Analysis | 1 Quick | 1 Quick | Disabled | Disabled |
| Research Mission | 1 Deep | Not included | 1 Deep | 1 Deep |
| Compare Screenshots | Not included | Not included | Not included | 1 Deep |
| Debug UI | Not included | Not included | 1 Deep | 1 Deep |
One row is worth explaining: Quick Analysis is available on Free and Starter but disabled on Pro and Business. That's on purpose. Pro and Business allocations are Deep credits, and every mode on those tiers uses the deeper analysis path. There's no "small credit" you could spend on a Quick summary. If you were on Pro and wanted a fast take, the answer is still to run a Research Mission (a Deep credit is still just one credit) or use one of the non-AI features like Extract Text.
What about the non-AI modes?
Below the four AI cards you'll see three more options: Extract Text, Extract Handwriting, and Scan Code. Two of those run locally on your machine (Extract Text uses Tesseract OCR, Scan Code uses ZXing for QR and barcodes) and cost nothing to run beyond having a paid plan. Extract Handwriting is the exception: it's a cloud call to a small vision model and charges one credit from your natural pool. Use them when you want raw text or data extraction and don't need narrative analysis.
Ready to pick a mode?
Install Vzlyze, capture something, and try the mode that matches your task. If you're not sure, start with Quick Analysis. It's the cheapest way to see what the model sees.
Get VZLyze for Chrome