Honest answers to the questions everyone asks first.

No. They are Monte Carlo drawsfrom a Gaussian copula joint fitted to the indication's real anchor cohort. For DMG, that anchor is 60 H3 K27M-mutant patients pooled from Gröbner (Nature 2018) and Petralia (Cell 2020), covering both the H3.3 and H3.1 variants. The simulated profiles are statistically representative draws, not records of real people. Read the methods.

n=60 for DMG. These are the numbers above which every confidence interval in the system is computed. We bootstrap over the real anchor n, not the simulated draw count. A 1,000-draw simulated cohort drawn from n=60 still reports intervals reflecting n=60. The draw count is a Monte Carlo resolution parameter, not an information parameter. We make this visible in every report header.

No. The methodology preprint is in draft and has not been posted yet; the methodology page is the current write-up. The preprint will be released under CC-BY 4.0 so anyone can read and check the methods. The implementation itself stays proprietary and the repository is private. Domain-expert review has not yet completed for any indication, and no customer-facing report ships until it does.

Three things make the reports decision-grade rather than vibes: (1) citations at source, every claim carries an inline pointer to a PMID, a DepMap cell line, a LINCS signature id, or a cohort statistic; (2) the Honesty Critic agent runs adversarially on every sentence and refuses ones without retrieved evidence; (3) every model call and every tool call writes an agent_logsrow you can audit afterwards. You don't have to take the report on faith; you can pull the audit log and check it.

The system surfaces that case directly. The 45 degree agreement plot in every report shows which gene pairs the copula preserved and which it didn't. The leave-one-out cross-validation harness reports per-gene Brier scores, including the cases where the joint copula loses to a marginal baseline at small n. We'd rather ship a finding that disagrees with a published headline (e.g. STAG2+TP53 in Ewing did not reach Bonferroni in our n=222 extract) than bury it.

Onkydra is a hypothesis generator and prioritisation engine. It is not: a diagnostic, a treatment recommender, a medical device, a substitute for wet-lab validation, a substitute for clinical judgement. Every generated report carries this line in the header and the footer. You can use Onkydra to decide which target to spend your next quarter's lab budget on; you cannot use it to decide what to put in a patient.

No. We are explicit about this in the product, the documentation, the Terms of Service. Clinical decisions go through clinicians, registered diagnostics, IRB-approved trials. The system is for pre-clinical and pre-trial prioritisation, not bedside care.

Eight rare and paediatric indications: H3 K27M diffuse midline glioma (DMG; the flagship), Ewing sarcoma, MPNST, ATRT, high-risk MYCN-amplified neuroblastoma, anaplastic thyroid cancer, rhabdomyosarcoma, KMT2A-rearranged paediatric AML. DMG is the only indication you can run today, and its library is draft pending DMG-expert review. Ewing is shelved and not selectable. The other six ship as their molecular profile libraries are curated and reviewed. Roadmap.

Yes. Any of the ~200 curated cancer-relevant proteins is in scope on the Researcher and Team subscriptions. Custom target classes outside that list (e.g. a novel scaffold or a non-cancer protein) are handled as a scoped engagement via a scoping call.

Yes. A custom indication engagement (from €15,000, setup plus maintenance) covers custom indication parameterisation: we curate the molecular profile library for your specific cancer, get it domain-expert reviewed, and ship it inside the workspace for your team. Email hello@onkydra.com to scope.

Researcher · €99/mo: one workspace, up to three live targets in the launch indication, memo export with citations, 14-day free trial, and weekly auto-rerun on ClinVar / DepMap / PubMed deltas (in build, not yet live). Team · €499/mo per team (not per seat): ten live targets today, with multi-seat access for up to five people, shared programme memory, drift alerts and scheduled reporting all in build and not yet live. The workspace is single-user at present; there is no team or seat model behind it. Engagements are scoped delivered work rather than a subscription: a target pilot from €1,500, a decision sprint from €10,000, a custom indication from €15,000, and a monitoring retainer from €750/mo.

Yes. 14 days on the Researcher tier, full feature access, no card required up front via Stripe Checkout. You can cancel any time inside the trial and you won't be charged. The Team subscription and the engagements do not offer a free trial; instead, drop us a line for a guided walkthrough.

You manage cancellations from the Stripe-hosted Customer Portal. One click. No retention dark patterns. Your data stays available for export via /workspace/api/me/export and can be deleted entirely via /workspace/api/me/delete per GDPR.

Annual pricing is live (17% off monthly, toggle on the pricing section). Foundation sponsorships are productised at /foundations: three tiers from €79 to €59 per seat / month starting at 20 seats, with a foundation dashboard, sponsor attribution on every report, and quarterly impact reporting. Current partner shape is ChadTough, DIPG Advocacy, and Cure JM-aligned programmes.

No. Onkydra v1 deliberately does not accept patient-identifiable data. The system operates entirely on derived statistical distributions from public datasets (cBioPortal, as the interface to the DKFZ pediatric pan-cancer and CPTAC pediatric brain studies, plus DepMap, LINCS and PubMed). Future versions may support bring-your-own clinical data inside a covered-entity BAA with Google Cloud; not in v1.

GET /workspace/api/me/export returns every row keyed to your account (runs, traces, cohort manifests, billing audit trail) as one JSON bundle with a lawful- retention notice covering Stripe-side invoice retention under GDPR Article 17(3)(b). POST /workspace/api/me/delete is a two-phase HMAC-confirmed flow that cancels any active Stripe subscription, deletes every uid-keyed Postgres row, and deletes the Firebase Auth user.

You do. Onkydra ships the report as your output; we don't claim derivative rights on the analysis your subscription paid for. The underlying open-data sources retain their licences (cBioPortal ODC-BY, DepMap academic, LINCS CC0, PubMed); we honour those upstream attributions in the citation layer.

Google Cloud Platform · europe-west2 (London). Postgres on Cloud SQL with multi-tenant row-level security; no client-side database access; per-secret IAM on the Cloud Run runtime service account; encryption at rest and in transit. The full stack is Google-Workspace-grade and meets the standard EU data-residency requirement.

withZeta.ai is the closest vertical match, same buyer profile, same "rare-cancer AI co-scientist" story. But the shape of the offering is different in three ways. First, indication depth: withZeta covers 438 cancer types via ontology (Orphanet, NCI Thesaurus, HPO); Onkydra ships per-indication anchor cohorts curated end-to-end (n=60 real DMG cases from DKFZ + CPTAC, live today and draft pending DMG-expert review; an n=222 Ewing anchor is built but shelved and not selectable, and six more indications are queued in the roadmap). Their ontology tells you the disease exists; our cohort lets you rank strata against real co-mutation structure. Second, model layer: withZeta's purpose-built modules are PredictBBB (blood-brain-barrier permeability) and ETHER0 (chemistry reasoning, Apache 2.0 from FutureHouse), chemistry-forward. Ours is transcriptomics-forward: a ridge baseline cross-checked against a mechanistic perturbation layer. That layer is a real CellOracle in-silico knockout over a DMG regulatory network, precomputed for 47 transcription factors and abstaining beyond them, with scPRINT-2 running as a labelled proxy until its serving endpoint deploys. Third, distribution: withZeta ships from Lantern Pharma, which also runs its own oncology pipeline; Onkydra is an independent tool vendor, not a drug developer.

No, and the honest answer takes four lines. First, vertical vs horizontal: Claude Science is a general-purpose workbench for any scientific discipline; Onkydra ships a pre-loaded molecular profile library anchored to a real pooled patient cohort (DMG today, n=60 from DKFZ and CPTAC; draft, pending DMG-expert review). Second, a built-in adversarial check: the Honesty Critic cross-checks the ridge baseline against a separate mechanistic perturbation layer on every report, and declares which layers ran as labelled heuristic proxies rather than live model calls. Third, the perturbation model is different: the NVIDIA BioNeMo bundle inside Claude Science includes Evo 2, Boltz-2, and OpenFold3 but not scPRINT-2, which is our default transcriptomic perturbation predictor. Fourth, evidence-size honesty: every Onkydra report bootstraps its CIs over the real anchor n (n=60 for DMG), not over the synthetic draws. Claude Science's reproducibility promise is code-level (every artefact is reproducible from the code snapshot); ours is statistical (every CI is honest about the real n). Different depth of field, same interaction quality bar.

Because a Claude Science user searching for “H3 K27M diffuse midline glioma” would spend the first 90 minutes of their session pulling and normalising the DKFZ and CPTAC pediatric K27M cohort by hand, another hour deciding whether Ledoit-Wolf shrinkage on Σ is the right regularisation for n=60 and 13 genes, and would then have to convince a peer reviewer their ad-hoc setup was sound. Onkydra ships all of that pre-curated, with the methodology documented and domain-expert review still outstanding. The €99/month is what those first three hours cost.

Not at the buyer level. Claude Science's launch customers are BMS (30,000 employees), Genentech, Novartis, Regeneron, AbbVie, Sanofi. Onkydra's customers are €99/mo PhD researchers, Seed-A biotech R&D teams (€499/mo per team), and founders commissioning scoped engagements. The buyer, price shelf, sales motion, and procurement gate are all different. Anthropic will also not compete with us on the eight-indication vertical because a horizontal Anthropic-scale platform will always add value on the axis with the largest customer bases, not on the eight indications that combined account for a few thousand new patients per year worldwide.

Yes, on the roadmap (task P73). A Claude Science user will be able to invoke Onkydra from inside their session: onkydra.list_indications(), onkydra.sample_cohort(indication, n), onkydra.run_target(indication, target), onkydra.check_target_across_indications(target). Distribution over ownership: Anthropic literally cannot compete with us on the eight rare-cancer indications if we're inside their session.

Complementary. Onkydra is built on the Gemini 3 family, so improvements in Gemini's scientific reasoning flow through to us directly. Gemini for Science bundles AlphaFold, AlphaGenome, and 30+ databases; Onkydra bundles a copula cohort engine, a ridge baseline, PubMed retrieval, and a mechanistic perturbation layer (CellOracle in-silico knockouts are real over a DMG regulatory network, precomputed for 47 transcription-factor regulators plus ACVR1 through an explicit ID1/ID3 bridge, and abstaining for every other target; scPRINT-2, DepMap and LINCS run as labelled proxies today), and rare-cancer indication libraries (one live today, DMG, draft pending DMG-expert review, with the rest on the roadmap). Different scope; we are a legitimate downstream use of Google's platform.

Founded by Faith Ogundimu, a cancer-genomics PhD researcher at the Royal College of Surgeons in Ireland (Dublin). The methodology has not yet been reviewed by an independent DMG domain expert. Independent review is a prerequisite before any customer-facing engagement ships, and no reviewer is named here until they have agreed to be. Onkydra is currently built and run by Faith alone. DMG is the only live indication today; the remaining seven rare-cancer indications are roadmap, not shipped.

Gemini 3.1 Pro (Planner, Honesty Critic, Workspace Writer); Gemini Flash family (Biology Engine); CellOracle for mechanistic perturbation, which is real: 47 transcription-factor knockouts precomputed against a DMG base gene regulatory network built from four scATAC-seq samples, served as an artifact, and abstaining outright for anything that is not a transcription-factor regulator in that network. scPRINT-2 (Cantini Lab) for transcriptomic perturbation, DepMap CERES/Chronos dependency anchoring and LINCS L1000 signature reversal run as labelled heuristic proxies today, with their serving endpoints not yet deployed; PubMed retrieval via Gemini text-embedding-005 embeddings in pgvector. AlphaGenome and MedGemma are wired for v2.

Not currently. The implementation, including the cohort sampler and the agent orchestration layer, is proprietary and the repository is private. What we intend to make public is the methodology: a preprint under CC-BY 4.0 describing the methods in enough detail to be checked and reproduced. That preprint is still in draft and has not been posted. This keeps the science checkable while letting the business sustain itself.

The anchor cohorts are built from public datasets and the methodology is documented, so the analysis is not a black box that disappears with us. Your export bundle (/workspace/api/me/export) gives you everything we hold for your account in one JSON file. Worst case, you lose the workspace UX; you don't lose the science or your history.