Guide
How to use BookGlints
A complete walkthrough of every feature — from your first scan to getting the most out of Auto-Scan.
Looking for troubleshooting, refunds, or a specific question answered fast? Check
Support instead — this page is a fuller walkthrough of how everything works.
1.Getting started
BookGlints starts in Native Local mode, so you can scan books without creating a BookGlints account or adding an API key. Semantic matching runs locally in your browser.
- Install BookGlints from the Chrome Web Store or Firefox Add-ons.
- Open a supported Goodreads, StoryGraph, or Royal Road book page, then open the BookGlints popup.
- Native Local is the recommended default. Add your flags or choose Auto-Scan, pick a scan depth, and scan.
- On the first Native Local scan, your browser may download the semantic model. The model is cached so later scans can start much faster.
- If you prefer cloud AI or stronger multilingual support, open Settings and switch to External AI. Choose Gemini, Groq, OpenRouter, or OpenAI, add your own API key, and save.
Tip: Native Local is private, fast, and best with English-language reviews. External AI is optional and useful when you want cloud-model behavior or stronger multilingual support.
2.My Flags vs. Auto-Scan
The popup has a toggle at the top with two modes:
- My Flags (free) — you tell BookGlints exactly what you're checking for, in your own words. It scans reviews and shows you the ones that genuinely match.
- Auto-Scan (one-time $9.99 unlock) — no flags needed. BookGlints reads the reviews and surfaces the recurring themes and patterns on its own, whether you know what to look for or not.
In Native Local mode, Auto-Scan finds recurring patterns using local semantic analysis and verified review evidence. External AI remains available if you prefer more open-ended cloud-model analysis.
Both modes share the same Scan Depth setting, work across Goodreads, StoryGraph, and Royal Road, and can run with Native Local or External AI.
3.Writing good flags
Add flags in Settings — one per line. BookGlints checks reviews semantically, not just for exact keyword matches, so you don't need to guess the exact wording reviewers might use.
- "Self-insert protagonist" also catches reviews that say "annoyingly perfect" or "everyone loves her for no reason"
- "Forced romance subplot" catches "the romance felt tacked on" or "didn't need the love triangle"
- Flags can be tropes, content warnings, pacing complaints, or just a vibe — "grimdark only," "no happy ending," "drags in the middle" all work
Tip: specific flags tend to work better than vague ones. "Chosen one who's good at everything" surfaces more precisely than just "overpowered."
4.Scan depths explained
Search Depth controls how much review coverage BookGlints gathers before analysis. The exact strategy adapts to Goodreads, StoryGraph, and Royal Road:
- Quick Scan — the fastest look at a book, using a smaller review sample.
- Focused Scan — prioritizes the strongest positive and negative reviews to surface clear praise and criticism.
- Full Scan — Recommended — scans the widest mix of reviews for the most complete picture. This is the default for new installs.
Full Scan is now fast enough to be the recommended everyday choice. Quick Scan remains useful for slower devices or when you only want a rapid first look. If you use External AI, deeper scans may also use more of your provider's allowance.
5.Batch scan & title import
Have a long to-read list? Open Batch Scan from the popup's header links.
- Choose My Flags or Auto-Scan at the top — same toggle as the main popup. Auto-Scan needs the one-time $9.99 unlock; My Flags uses whatever flags you've set in Settings and is always free.
- Paste book titles, one per line, and click Scan All — each is searched and scanned automatically.
- Or import titles from a supported list you already have open: Goodreads, StoryGraph, or your Royal Road Read Later list. BookGlints imports the titles currently loaded in that tab.
- Batch mode always uses Quick Scan for each book — quick and consistent, with no setup needed per title. Use the main popup for Focused or Full Scan on books you're seriously considering.
- With Native Local, batch analysis stays on your device. With External AI, each book can make a separate request to your chosen provider, so keep that provider's rate limits in mind.
6.Find another book by title
Don't want to open the book's page first? In the popup, use Find Another Book. Choose Goodreads, StoryGraph, or Royal Road, type a title, then pick the right result. BookGlints can scan it without you navigating away from the page you're already on.
7.History
Every scan — single or batch — is logged automatically. Open History from the popup's header to browse past scans, tagged with whether you used My Flags or Auto-Scan. It's a read-only log, kept to the most recent 200 scans, stored locally in your browser.
8.Understanding your results
- Prevalence signal — each match shows how many reviews mention it out of how many were scanned, so you can tell if it's a real pattern or a single reviewer's opinion.
- Verified excerpts — every quote shown is checked against the actual review text before it's displayed. If it can't be verified as a genuine excerpt, it's dropped rather than shown — nothing you see is AI-generated summary, only real reviewer words.
- Auto-Scan themes — grouped by category (Characters, Plot & Tropes, Pacing, Writing Style, Content & Warnings, Ending, Other) with a sentiment label (positive, negative, or mixed).
9.Language support
The BookGlints interface is available in English, Arabic, German, Spanish, French, Brazilian Portuguese, and Simplified Chinese.
Native Local semantic matching is optimized for English-language reviews. If you need stronger multilingual review analysis, switch to External AI; multilingual quality then depends on the provider and model you choose.