- Full 4-chapter curriculum
- Rotation curve simulator
- Command Vessel dashboard (demo feed)
- Open theory repositories
- Everything in Free Explorer
- Full 10-lesson course: Micro to Macro (Python labs included)
- Cloud optimization engine on live SPARC data
- Saved fits + dataset versioning
- Hosted telemetry endpoint for the Vessel
- Research community + methods workshops
Marked ◷ = in build — everything else on this page is live, including the engine: hosted fits run the exact code behind the published SPARC v3.0 sweep, certified reproducible to the last digit. What remains is checkout, connecting tonight; founding researchers lock the $30 rate for life.
Your account
Accounts are live now — create yours with a passwordless magic link. Checkout opens when payment processing lands; notify-list members lock the $30 founding rate for life.
What the engine runs
The same pipeline behind the published results — automated, global, and honest about uncertainty:
Ingest
Ingest the official SPARC release with full provenance — dataset version and file hash pinned to every fit.
Optimize
Fit each galaxy with the two-parameter QAG law — no dark-matter halos, no hidden knobs — and score Newton + MOND on the same points.
Challenge
Every result ships with its Newton and MOND baselines on identical data, in public — built to try to break the framework, not to confirm it.
Latest validation run
Full SPARC sweep, corrected pipeline (Oct 18, 2026) — 175 galaxies, 3,391 data points, per-galaxy least-squares fits, Newton + MOND baselines on identical data. Reduced χ² ≈ 1.0 is the ideal; SPARC's tiny formal errors keep every model above it, so read this head-to-head:
5.55 (median 1.42)
Reduced χ² across 175 galaxies. Wins 167/175 vs MOND, 170/175 vs Newton.
56.55
Standard MOND, fixed a₀ = 1.2×10⁻¹⁰ m/s² — 10× worse than QAG.
616.74
Baryons only, no dark matter — 111× worse than QAG.
ΔBIC (MOND−QAG) = 172,054: decisive. Methods, per-galaxy tables, and the audit that caught the old circular formula — all public.
Cross-framework validation — October 20, 2026
The pipeline then went head-to-head with a second published framework (Stålhammar's Emergent Scalar Field, 2026): its Equation 8 re-implemented from the paper and replicated exactly on its home ground (SPARC 0.227 vs published 0.228 dex; X-COP γp = 0.69, LOCO 0.072), both frameworks cross-run under both scoring protocols, and QAG's new Cluster Law taken to X-COP, eROSITA, and the Bullet Cluster lensing map. Reports 1–8, all public.
0.166 dex
QAG pointwise law on X-COP hydrostatics — the calibration sample for βg. ESF's γp = 0.69 replicated exactly on the same data; MOND 0.138; GR 0.800.
0.206 dex
QAG pointwise law on eROSITA (3,916 points) with its constant frozen from X-COP — radially flat. ESF frozen: 0.441 (its published failure, replicated); MOND 0.514; GR 1.035.
0.83 vs 0.85
Lensing peak-height ratio, model vs the actual Clowe+06 map. The affinity density puts the subcluster peak on the galaxies (20–80 kpc; observed 43 kpc); baryons alone produce no peak at all.
β reconciliation
The cluster law's integral form prefers a different constant on each system (0.005 / 0.028 / 0.225) while its pointwise form holds one (βg ≈ 8×10¹¹ s²/m). Deriving the field equation that unifies them is the live research question — memo public.
For institutions
Teaching QAG in a course or running a journal club? Department licenses with classroom workspaces and instructor tooling are on the roadmap — reach out and we'll build it with you.