FoldLens
iPad-first AlphaFold Database browser with honest pLDDT confidence display and on-device Apple Intelligence residue explanations. The lens through which everything else flows.
- Viewer
40 apps across the portfolio — 35 usable right now in your browser, no install. All free forever, all on a path to open source under Apache 2.0, all built by Fold Commons, a non-profit organisation. None of them collect anything about you.
Selecting a filter updates the list below.
Showing all 40 apps
iPad-first AlphaFold Database browser with honest pLDDT confidence display and on-device Apple Intelligence residue explanations. The lens through which everything else flows.
The K-12 (ages 9–14) sibling of FoldLens: the same AlphaFold Database content in a tier-appropriate voice, NGSS-aligned, COPPA-safe by design. Pick a protein and see its real 3D shape in the browser.
A reader over the Fold Commons research archive — search and read deep reports on AlphaFold and AI for biology, entirely in your browser. The native app adds on-device Apple-Intelligence grounded summaries.
AlphaFold vs Experiment: superpose an experimental structure (PDB) and AlphaFold’s prediction (AFDB) client-side for an honest RMSD diff and per-residue divergence profile.
A binding-pocket finder: enter a UniProt accession or PDB id and it finds candidate cavities in your browser, ranked by a transparent geometric heuristic. Geometry, not a druggability predictor.
Find the epitope/paratope interface of an antibody–antigen complex from geometry, clustering the paratope into CDR-like contact loops. A geometric contact map, not an experimentally-mapped epitope.
A phenotype → causal-gene → protein-structure explorer over well-established monogenic conditions, with a live AFDB render and key-residue highlight. Not a predictor or diagnostic.
The structural-biology Nobel Prizes as an interactive timeline, with verbatim citations, primary-paper DOIs, and 3D deep-links into the viewer.
An honest scoreboard of every CASP experiment (1994→2024) built from real published results — no invented GDT_TS, era filters, and a same-origin trajectory chart.
A sourced tracker of AI-in-clinical-trials milestones, every entry cited, with a strict no-causation posture and honest coverage gauge.
Public AI-biotech capital flows (IPOs, partnerships, funding, M&A) — every figure cited to a primary source, with an honest coverage gauge and an annual-totals chart.
Landmark closed-loop / self-driving-lab discovery runs, each a five-stage loop cited to its primary paper by DOI, with a hand-drawn closed-loop diagram.
Guided, textbook-sourced tours of eight cell types with ordered stops, honest coverage, and step-through — a function-first reader for how cells are built.
A COPPA-safe amino-acid sorting game — sort each of the 20 standard amino acids into its property class. No accounts, no outbound links, nothing collected.
A COPPA-safe science-writing game — call each protein claim mechanism or purpose, with no clock, and rewrite teleology as honest physics.
Predict what gets through each real membrane channel (aquaporin-1, GLUT4, CFTR) before the reveal — scored on your prediction, not luck, with a live AlphaFold render.
Set the ATP level and predict a molecular motor’s drift — a biased-random-walk game about why motors ratchet through jitter rather than stride.
Judge a crisp AlphaFold model’s domain arrangement, then reveal the PAE — your call scored live from the fetched predicted-aligned-error matrix.
Tag a protein’s low-confidence stretches Broken or Working — the reveal proves Working from live UniProt functional annotations.
Draw a real one-letter variant and call its impact before the reveal — a calibration game about why most mutations are quiet and a few are not.
Paint a Firm-or-Fuzzy prior on two crisp domains, then reveal the real PAE — a gentle confidence game (ages 9–14), scored live from AlphaFold.
Pick what a single-chain AlphaFold model leaves out — proven live from UniProt cofactor and subunit annotations.
Predict whether a pocket is a rigid keyhole or reshapes on contact — an induced-fit game with live enzyme renders (ages 9–14).
Set the cooperativity between hemoglobin’s four subunits and match the real oxygen curve — a game about why one O₂ makes the others bind more readily, but not all-or-none (ages 15–18).
How much does one protein change across the tree of life? Each organism’s version is the reviewed UniProt entry for its gene, compared with ours letter by letter and by predicted shape. Plus ZooFold: call whose version is closest to ours.
A visionOS-native, protein-scale spatial viewer — hold a structure in your hands, walk around a complex. Spatial depth a browser can’t match, so there is no web edition.
A co-folding viewer for protein–ligand complexes with transparent pose scoring. Planned once the co-folding output format stabilises.
A conformational-ensemble timeline for AlphaFlow / DiG / DynaFold output. Planned once the ensemble format stabilises.
A Molecule-of-the-Month-style guided reader with a live Mol* render. In early design.
A viewer for viral capsids and glycoproteins — large assemblies rendered honestly. In early design.
An accessibility-first listening game: AlphaFold confidence sonified along the chain as pitch. Match the sound to the protein and learn that low confidence is not the same as unimportant (ages 9–14).
Drag AlphaFold’s confidence lens and predict where each region vanishes — a game that shows a faded region is usually flexible and functional, not a mistake (ages 9–14).
Carve a protein into its folded domains, then use AlphaFold’s PAE to judge how much to trust the way the domains sit relative to each other (ages 15–18).
Paint which parts of a protein tolerate change, then reveal the real AlphaMissense heatmap — some proteins shrug off most changes, others are sensitive almost everywhere (ages 15–18).
See one subunit, bet which parts get buried against its partner, then load the real assembly and score your guess against the true interface — many proteins don’t work alone (ages 9–14).
You are the renderer: choose which parts of a protein to draw in firm ink and which to draw sketchy. Sketchy is how you honestly draw a flexible, functional region (ages 9–14).
Compare two spliced isoforms of one gene and pick the exon that explains the difference — splicing makes different chains, not one chain morphing into several shapes (ages 15–18).
A space-filling model looks like a solid pile of marbles, but atoms jiggle — bet which regions hold still and which smear, then watch the mobile ones jitter (ages 9–14).
A suggested learning ladder: nineteen rungs from “a cell has parts” to “why prediction is trustworthy”, with two age-band entry points. Each rung links to a free tool, and the prerequisite arrows show what tends to make the next step easier.
Read the real AlphaFold confidence numbers off human haemoglobin: guess how sure the model is for a region, then see its actual per-residue pLDDT and pairwise PAE — computed live from the model — and say it plainly: a prediction, not an experiment (ages 15–18).
No apps match those filters.
Each app sits on top of FoldKit, a shared open-source Swift library — the AlphaFold Database client, the Metal rendering pipeline, the reusable SwiftUI chrome. We don't copy-paste; we reuse. Each web tool is a zero-install re-implementation that fetches public data directly to your browser (or runs entirely offline), never through a server of ours. All source code is private during build-out and goes public at v1.0 under Apache 2.0; the research archive content is CC BY 4.0.