Scan a barcode or photograph a whole shelf. Our Library catalogues every title, catches double-buys, tracks what everyone is reading, and helps you choose what comes next β on private, Australian-operated infrastructure.
From the first barcode scan to the next great recommendation, every feature is designed to make a large shared collection feel simple.
π·
Scan in seconds
Use a phone camera to scan an ISBN, or photograph an entire shelf and let vision AI identify the spines.
π
Stop double-buying
See an immediate warning when a book is already owned, including who has it and when it was added.
π₯
One home, separate libraries
Every family member gets a private login, personal collection, reading history, ratings, and reviews.
β¨
Ideas that learn
Get personalised reading suggestions shaped by your own books, wishlists, and βnot interestedβ feedback.
π
Complete every series
Automatically group series, spot missing volumes, and add the next title to a wishlist in one tap.
π
Find the best place to buy
Compare Australian stores, watch wishlist prices, and jump to nearby library holdings through WorldCat.
Powered by insta.host
AI-optimized technology. Kept close to home.
The app runs on private, Australian-engineered infrastructure: a dependency-free Node.js service, an on-premises SQLite database, and carefully scoped AI assistance for the jobs it does best.
insta.host managedNode.js 22SQLite WALTrueNAS SCALEAzure AI FoundryLibraryThingGPT-5.6 SolCloudflare Tunnel
π±
Mobile-first web appBarcode scanning, shelf photography, library browsing, and reading tracking from any modern browser.
β
π
Private insta.host accessManaged HTTPS access through an outbound-only secure tunnel, without exposing the home network directly.
β
π¦πΊ
Australian-operated application and dataNode.js runs on insta.host-managed hardware; the family catalogue stays in a local SQLite database under sovereign control.
β
π€
AI and trusted book sourcesAzure AI Foundry reads shelves and creates suggestions. GPT-5.6 Sol conservatively fills metadata gaps after Booko, OpenLibrary, and Google Books.
π‘οΈ
Private by design. Operated by insta.host.
Passwords are securely hashed, sessions use protected cookies, and each person's library and recommendation history remain separate on privately managed infrastructure.
New scans are saved only to the selected person's library.
Point the camera at a book's barcode
Shelf matches stay private to the selected person's library.
π· Scan a whole shelf
Use a clear photo of a shelf (one row at a time works best) β take one now, or pick one you already have. The AI reads the spines, then you confirm what to add.
Recommendations and feedback stay scoped to the selected person's library.
β¨ Ideas for your next read
Personalised picks based on your own library.
π Find a book you can only describe
For when you remember the story but not the title β βthe one about a beekeeper in wartime Franceβ. Searches LibraryThing's Talpa. The household shares a small daily allowance, so it only runs when you ask.
π₯ People
Create a private login for each reader. Every person keeps a separate library, reading history, ratings, reviews, and suggestions.
Add a person
Import QBD purchase history
Paste a saved QBD order-history page or order confirmation email. The importer extracts ISBNs/product links and adds them to the selected library without storing QBD credentials.
π Find duplicate purchases
Scans a library for books bought twice (matched by ISBN, by edition when LibraryThing knows the book, or by title when there's no barcode) and flags every extra copy as a duplicate to sell or donate β grounded on whichever copy was scanned first.
π ISBN health check
Reviews the barcodes already catalogued. Flags entries that are not valid ISBNs, and editions from the wrong market (a translated edition attached to an English copy) β both quietly break double-buy warnings, price matching and store links.
πΌοΈ Fill in missing book details
Looks up covers, synopses, publishers and page counts for books that are missing them, and stores each cover on this server so it keeps working and stops telling other sites what you read.
π€ Export your library
Download every book, rating, review and reading date as a file you can keep, open in a spreadsheet, or move elsewhere.
π‘ Book data sources
Shows whether each catalogue BookLib looks books up in is answering. A source that stops responding used to fail silently.
A private family book library. Scan a barcode or snap a photo of a shelf and it's catalogued
automatically β cover, description, publisher and price comparisons across major retailers.
Track what's owned, wishlisted or read; rate and review; get personalised "read next" picks
that learn from what you like; and catch accidental double-buys before they happen.
Under the hood
Hosting: runs entirely on a home server, not a public cloud β your library lives in a local database on your own hardware, not a third party's.
Access: reachable over HTTPS from anywhere through an outbound-only secure tunnel β no inbound ports are opened on the home network.
AI β vision & suggestions: shelf photos are read, and personalised suggestions are written, by a GPT-family model hosted on Azure AI Foundry through a private gateway.
AI β metadata assist: when the public book databases below don't have a synopsis, publisher or page count for a title, GPT-5.6 Sol conservatively fills the gap β it's only ever asked to describe a book a real catalogue already confirmed exists, never to guess one from scratch.
Book data & pricing: covers, descriptions and prices come from Booko, OpenLibrary, Google Books and LibraryThing, cross-referenced against Australian retailers (QBD, Big W, Dymocks, Amazon, Books Ever After) and Barnes & Noble in the US, and WorldCat for library holdings.
Editions & series: LibraryThing tells the library which ISBNs are the same book, so a second copy in a different edition is still caught as a double-buy, and which series a book belongs to. Used non-commercially, within LibraryThing's daily request allowance.
Accounts & privacy: each family member has their own login β libraries, ratings, reviews and suggestion feedback are kept separate per person and never shared or merged.
π¬ Send feedback
Found something broken, or thought of something the library ought to do? It is saved in
this library's own database on the home server β nothing is sent to anyone else.