Library

The library is 48 worked programs, every one tagged on the same five axes so a filter never lies about what it hid.

#Two ways in

• The gallery — open itin the lab from the Explorer's empty state, the Library rail's Browse the gallery button, or ⌘⇧P → Library gallery. A full-page room grouped into five shelves. / focuses search, typing filters live, ↓ steps into the grid, arrows walk cards (Down crosses into the next shelf), Enter opens, Escape closes.
• The Library rail (⌘K) — for someone mid-edit who wants one thing. Search by subject, not just title: "entangle" finds the Bell state. Filters: All / Runs here / Reading, level (intro / core / advanced), kind (primitive / algorithm / application), domain (foundations, search, optimization, chemistry, machine learning, cryptography, finance, logistics…). The palette lists every item by name too: ⌘⇧P → Library: <title>.

#The five shelves

1. Start with these — four items to show someone in their first minute: each runs in your browser, draws what it is doing, and quotes what it produced when last run. Every card here also lives on its own shelf below.
2. Business cases — the cards a buyer asks about, curated at the top. The first is a decision aid: three problems solved both ways, one table, and the rule of thumb the rest of the shelf supports. Then healthcare (a screening classifier on 100 real biopsies, kidney exchange), vision and inspection (quanvolution vs a classical filter on real digits, edge maps), pharma (VQE on H₂), finance (credit-risk queue, fraud screening, portfolio, option pricing, VaR), logistics, energy, telecom, security. Each card carries a one-line hook — industry · the decision it informs · the verdict the program prints — and the verdict is honest: most of these still lose or tie against the free classical tool, and the card says so. Every business card also lives on its ordinary shelf below.
3. Applications — algorithms pointed at problems an industry would recognise. Every card runs in your browser as written.
4. Algorithms — the named results the field is built on: Grover, QAOA, teleportation, error correction.
5. Building blocks — superposition, entanglement, interference. Start here if the notation is new.

#The honest labels

• Runs here, free — executes on the in-browser simulator; no network, no credits.
• Reading only — for an item that cannot execute here. No item carries it today: everything runs.
• A chip icon marks items that ship a .qasm the Run rail can submit to hardware — the point is the gap between a noiseless prediction and a device.
• Runnable items quote what was observed when run during development; the test suite re-runs each shipped circuit on the simulator against those numbers on every deploy, and syntax-checks each Python program.
• Items built on public data (the health, credit-risk and vision cases use UCI datasets, CC BY 4.0) fork with a data.csv, a README stating the row rule and what changed, and a Data in this item block with the creators, DOI and licence — the attribution travels with your copy. Where data is generated, the item says so first.

#Forking

Open an item: summary, what it teaches, level and domain, the files with a preview. Fork into a project copies the files into a new project of yours, opens the item's entry file, and closes the gallery. The copy is yours — edit it, share it, submit its .qasm to hardware. Forking again makes another copy; the library itself never changes.

Related: Getting started · For hackathons