RESEARCH USE ONLY

Plans for the teams taking on rare cancers.

The same engine behind every tier. What changes between them is how much of it you can point at once, who is allowed to use it, and whether a named person stands behind the answer. A hypothesis generator and prioritisation engine, not a diagnostic.

PRICING

One ladder, from a free benchmark to a named programme.

An academic lab pays per lab, with no seat count, because a research group loses and gains people by design. A commercial buyer pays per seat, or per team with seats included. A programme is one named indication or target, agreed in a project plan, time-boxed, and expandable only by amendment. Research use only: a hypothesis generator and prioritisation engine, not a diagnostic and not a medical device.

Provisional pricing. These numbers are set from published comparables and are not settled. Expect them to move.

Checkout is not open, so nothing here can be bought today and no card is taken anywhere on this site.

Free

Anyone deciding whether Onkydra can answer their question at all.

€0free forever

One seat.

Join the waitlist

Academic lab

One named academic lab, where no for-profit owns the intellectual property in the work.

€300per lab

€249 a month billed yearly.

No seat count.

Join the waitlist

Commercial seat

One named person at a company, or an academic on a programme whose IP a for-profit owns.

€800per seat

€664 a month billed yearly.

One seat.

Join the waitlist

Commercial team

A rare-cancer R&D team running more than one candidate at a time.

€1,600per team

€1,328 a month billed yearly.

3 seats included. €300 a month per extra seat.

Join the waitlist

Programme

A team betting the next twelve to eighteen months on one target or one indication.

Quotedfrom €40,000 per year

Scope a programme

Compare every tier, line by line →Join the waitlist

What changes at each step

This lists what DIFFERS between two adjacent tiers, in both directions: a commercial seat covers one person where an academic lab covers a whole group, and that shows up here as the reduction it is.

Free → Academic lab

Indications in scopeRead-only, on the public benchmark → H3 K27M DMG, within its stated scope
Runs of your ownNot included → Included
Daily run ceiling4 a day → 5 a day
Targets under watchNone → 3 targets
Mechanistic in-silico knockoutNot included → Included

and 8 more differences in the table below.

Academic lab → Commercial seat

People includedUnlimited, in one named lab → 1 named user
Commercial use of the outputNot included → Included
Where questions goEmail → Email, plus an onboarding walkthrough
Invoice and purchase-order billingNot included → Included
Data-processing agreementTemplate published → Counter-signed

and 1 more difference in the table below.

Commercial seat → Commercial team

Daily run ceiling5 a day → 15 a day
Targets under watch3 targets → 10 targets
People included1 named user → 3, then a flat rate for each extra
Shared programme memoryNot included → In build, not live
Where questions goEmail, plus an onboarding walkthrough → Email, onboarding, and a quarterly programme review

and 2 more differences in the table below.

Commercial team → Programme

Indications in scopeH3 K27M DMG, within its stated scope → DMG, plus one named indication or target curated for you
Runs of your ownIncluded → Agreed in the project plan
Daily run ceiling15 a day → Agreed in the project plan
Targets under watch10 targets → Agreed in the project plan
An indication or target curated for youNot included → Included

and 7 more differences in the table below.

Full comparison

Every line, on every tier.

Monthly rates below. Annual billing is 17% off on every tier that has a monthly rate, on the toggle above. Nothing on this page can be bought today.

Onkydra pricing tiers compared, feature by feature. Prices are provisional and nothing can be bought yet.
Feature	Free€0free forever	Academic lab€300per lab	Commercial seat€800per seat	Commercial team€1,600per team	ProgrammeQuotedfrom €40,000 per year
What you can analyse
Indications in scopeOne indication is wired today: H3 K27M diffuse midline glioma, on an n=60 anchor pooled from DKFZ and CPTAC via cBioPortal, with 13 driver features and no treatment-response or survival layer. It runs within that stated scope. Every other indication is on the roadmap and the coverage gate refuses it rather than guessing.	Read-only, on the public benchmark	H3 K27M DMG, within its stated scope	H3 K27M DMG, within its stated scope	H3 K27M DMG, within its stated scope	DMG, plus one named indication or target curated for you
Coverage check before you commitAsk whether Onkydra can answer about a target in an indication and get yes or no with the reason for a no. It is the same gate the run path uses, so a refusal here is the refusal you would get in a run. Available without paying anything, deliberately.	Included	Included	Included	Included	Included
Runs of your ownThe free tier grants zero compute. That is the step up to any paid tier and there is no softer way to say it.	Not included	Included	Included	Included	Agreed in the project plan
Daily run ceilingA ceiling on spend, not a price basis. Nothing is sold by the run: there is no top-up, no credit pack and no consumable anywhere in checkout, and an exhausted ceiling sells you nothing and waits for the 24-hour window to roll. An abstention is metered on its own counter at three times the allowance, so an honest refusal is never the expensive answer.	4 a day	5 a day	5 a day	15 a day	Agreed in the project plan
Targets under watchOne slot is one target watched inside one investigation, which is a standing commitment to check evidence on your behalf. The watching itself, weekly re-runs and drift alerts when ClinVar, the dependency rows or the literature move, is in build and is live on no tier today.	None	3 targets	3 targets	10 targets	Agreed in the project plan
Mechanistic in-silico knockoutReal CellOracle knockouts over a DMG regulatory network, served for 45 transcription factors and for ACVR1 through an explicitly labelled bridge to its downstream factors. It abstains for everything else, which is why CDK7, MEK1, PI3K, ONC201 and MENIN get no mechanistic prediction. An uncalibrated ordering signal and never a probability. Identical on every paid tier.	Not included	Included	Included	Included	Included
Measured screen evidencePooled CRISPR dependency in the six DMG lines from the Childhood Cancer Model Atlas deposit, and 1,430 transcription factors measured in DIPG17 from GSE267415. A screen measures what a cell needs in order to proliferate, which is a different question from what controls its state, and every surface serving one says so. Identical on every paid tier.	Not included	Included	Included	Included	Included
Tumour-tile morphology descriptionMedGemma describes the morphology of an H&E tile: cellularity, nuclear features, architecture. It is asked for a description and never for a diagnosis, a grade or a subtype, and the prompt is fixed server-side so it cannot be asked for one.	Not included	Included	Included	Included	Included
An indication or target curated for youCurating an indication means building and verifying its anchor cohort, its regulatory network and its screen evidence, then stating what is missing. It is scoped work with a project plan behind it, which is why it is not a monthly add-on at any subscription tier.	Not included	Not included	Not included	Not included	Included
Evidence and provenance
Every claim resolved to a citationPMIDs and DOIs are resolved against NCBI and CrossRef at run time, and a citation that does not resolve does not reach the output. This is what stops a fabricated reference, and it runs on every tier including the sample report.	Included	Included	Included	Included	Included
Retracted sources held, not citedA cited paper that has been retracted holds the run rather than appearing with a footnote. This has fired on a real run.	Included	Included	Included	Included	Included
Evidence class declared on every layerMeasured result, model run, stated rule, or labelled placeholder, named per layer and never upgraded by the report writer. Negative, null and uninterpretable results are preserved rather than dropped.	Included	Included	Included	Included	Included
Abstention when the target is out of scopeOnkydra refuses and says why, rather than producing a confident answer about a target it has no grounded evidence for. A refused run does not consume the daily run ceiling.	Included	Included	Included	Included	Included
Full artefact exportThe complete run artefact, including inputs, tool calls, evidence classes and citations, exportable as JSON so a result can leave the product and be checked by somebody who does not have an account.	Not included	Included	Included	Included	Included
Intervals resampled over the real anchorEvery interval is bootstrapped over the 60 real cases, never over the simulated draws. A simulated co-mutation profile is a Monte Carlo draw from one fitted model and is never a patient, which is also why nothing here is ever priced by how many of them a run makes.	Included	Included	Included	Included	Included
The experiment
Prediction frozen before the result is knownThe prediction ledger writes a claim, its evidence and its date before anybody knows the answer, so it cannot be edited to fit the outcome afterwards. The ledger is real and no prediction has been resolved yet, because no result has come back.	Not included	Included	Included	Included	Included
The experiment designed with usOn the subscription tiers you decide what to test and you carry the interpretation. On a programme the experiment is specified in the project plan with us before it runs, which is the difference between a tool you operate and an answer somebody stands behind.	Not included	Not included	Not included	Not included	Included
Observed result filed against the predictionThe result that comes back goes into the immutable record against the prediction that preceded it, whether it agrees or not. Resolution is built and has never run, because nobody has returned a result yet, and this table will say so until one does.	Not included	Self-filed	Self-filed	Self-filed	Filed with us, against the project plan
A named person accountable for the answerThe credence-goods literature is the honest framing here: you cannot verify this kind of answer even after delivery, and the two things that make such a market work are liability and verifiability. The subscription tiers sell capability, so you carry the risk of having used it wrongly. The programme tier sells somebody standing behind the answer. Different jobs, which is what makes the multiple coherent instead of insulting.	Not included	Not included	Not included	Not included	Included
Collaboration and seats
People includedThe academic tier is priced per lab rather than per person because a research group of three to fifteen churns people by design, and two 10-K filings in this sector state plainly that seat pricing shrinks when a customer loses staff. Charging a lab for that is charging it for being a lab.	1	Unlimited, in one named lab	1 named user	3, then a flat rate for each extra	Agreed in the project plan
Shared programme memoryIN BUILD, LIVE ON NO TIER. The workspace is single-user today and there is no team or seat model behind it. It is listed because it is what the team tier is for, and it is marked because nobody should buy it expecting it this month.	Not included	Not included	Not included	In build, not live	In build, not live
Commercial use of the outputThe academic tier is fenced on employer and IP ownership rather than on an email domain. Work inside your employment at a for-profit, and contract research where a for-profit owns the intellectual property, both need a commercial licence even when the person doing it sits in a university.	Not included	Not included	Included	Included	Included
Support
Where questions goAll customer-facing founder time comes out of one weekly ceiling, and support and scientific authoring draw on the same hours. The ceiling is a stated decision rather than a measurement, because there are no customers to measure. It is published so that a promise this product cannot keep is visible before you buy rather than after.	Public docs and the capability register	Email	Email, plus an onboarding walkthrough	Email, onboarding, and a quarterly programme review	A named contact, inside the project plan
Founder time is capped, and the cap is publishedFive hours a week of customer-facing founder time, in total, across every customer. A ceiling and not a target, and it exists because the alternative failure is invisible: hours quietly leaving the research one reasonable half-hour at a time.	Included	Included	Included	Included	Included
Procurement
Card, monthly, cancel anytimeMonthly billing keeps every transaction under the university purchasing-card limits observed across the institutions checked ($2,500 to $5,000) and under both stated purchase-order floors, so the purchase never has to touch procurement. Watch one clause if you are at a US university: at least one bars services from the card entirely, at any amount.	Included	Included	Included	Included	Not included
Annual billing at a discountSeventeen percent off the monthly rate, billed yearly. It clears the academic purchase-order floor on purpose, for a buyer who would rather pass through procurement once a year than expense a card charge every month.	Not included	Included	Included	Included	Annual by default
Invoice and purchase-order billingThe honest upgrade gate for a small commercial buyer. It is not a feature anybody wants: it is something their organisation demands, which is exactly why single sign-on works as a gate for larger vendors and cannot work here, since a solo PI and a three-person team have no identity provider to ask.	Not included	Not included	Included	Included	Included
Data-processing agreementThe standard template and the sub-processor register are published and readable by anyone. A counter-signed agreement, and a customer-specific annex where one is needed, come with a commercial licence.	Template published	Template published	Counter-signed	Counter-signed, with an annex where needed	Counter-signed, with an annex where needed
Security questionnaire completedWe complete your questionnaire, including the HECVAT that several US universities require before any software is acquired, free items and renewals included. None has been completed yet, because no customer has asked for one.	Not included	Not included	On request	Included	Included
Validated-system documentationNOT OFFERED AT ANY TIER, and it is in this table so that nobody assumes it comes with the largest one. Onkydra is a hypothesis generator and a prioritisation engine, research use only, and it makes no 21 CFR part 11 or EU Annex 11 claim. If your process needs a validated system, this is not one.	Not included	Not included	Not included	Not included	Not included
Research use onlyNot a diagnostic, not a medical device, not a treatment recommender, and not a substitute for wet-lab validation. This does not change at any price, and a tier that relaxed it would be selling something Onkydra is not.	Included	Included	Included	Included	Included

What you can analyse

Indications in scope
One indication is wired today: H3 K27M diffuse midline glioma, on an n=60 anchor pooled from DKFZ and CPTAC via cBioPortal, with 13 driver features and no treatment-response or survival layer. It runs within that stated scope. Every other indication is on the roadmap and the coverage gate refuses it rather than guessing.
Free
Read-only, on the public benchmark
Academic lab
H3 K27M DMG, within its stated scope
Commercial seat
H3 K27M DMG, within its stated scope
Commercial team
H3 K27M DMG, within its stated scope
Programme
DMG, plus one named indication or target curated for you
Coverage check before you commit
Ask whether Onkydra can answer about a target in an indication and get yes or no with the reason for a no. It is the same gate the run path uses, so a refusal here is the refusal you would get in a run. Available without paying anything, deliberately.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Runs of your own
The free tier grants zero compute. That is the step up to any paid tier and there is no softer way to say it.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Agreed in the project plan
Daily run ceiling
A ceiling on spend, not a price basis. Nothing is sold by the run: there is no top-up, no credit pack and no consumable anywhere in checkout, and an exhausted ceiling sells you nothing and waits for the 24-hour window to roll. An abstention is metered on its own counter at three times the allowance, so an honest refusal is never the expensive answer.
Free
4 a day
Academic lab
5 a day
Commercial seat
5 a day
Commercial team
15 a day
Programme
Agreed in the project plan
Targets under watch
One slot is one target watched inside one investigation, which is a standing commitment to check evidence on your behalf. The watching itself, weekly re-runs and drift alerts when ClinVar, the dependency rows or the literature move, is in build and is live on no tier today.
Free
None
Academic lab
3 targets
Commercial seat
3 targets
Commercial team
10 targets
Programme
Agreed in the project plan
Mechanistic in-silico knockout
Real CellOracle knockouts over a DMG regulatory network, served for 45 transcription factors and for ACVR1 through an explicitly labelled bridge to its downstream factors. It abstains for everything else, which is why CDK7, MEK1, PI3K, ONC201 and MENIN get no mechanistic prediction. An uncalibrated ordering signal and never a probability. Identical on every paid tier.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Measured screen evidence
Pooled CRISPR dependency in the six DMG lines from the Childhood Cancer Model Atlas deposit, and 1,430 transcription factors measured in DIPG17 from GSE267415. A screen measures what a cell needs in order to proliferate, which is a different question from what controls its state, and every surface serving one says so. Identical on every paid tier.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Tumour-tile morphology description
MedGemma describes the morphology of an H&E tile: cellularity, nuclear features, architecture. It is asked for a description and never for a diagnosis, a grade or a subtype, and the prompt is fixed server-side so it cannot be asked for one.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
An indication or target curated for you
Curating an indication means building and verifying its anchor cohort, its regulatory network and its screen evidence, then stating what is missing. It is scoped work with a project plan behind it, which is why it is not a monthly add-on at any subscription tier.
Free
Not included
Academic lab
Not included
Commercial seat
Not included
Commercial team
Not included
Programme
Included

Evidence and provenance

Every claim resolved to a citation
PMIDs and DOIs are resolved against NCBI and CrossRef at run time, and a citation that does not resolve does not reach the output. This is what stops a fabricated reference, and it runs on every tier including the sample report.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Retracted sources held, not cited
A cited paper that has been retracted holds the run rather than appearing with a footnote. This has fired on a real run.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Evidence class declared on every layer
Measured result, model run, stated rule, or labelled placeholder, named per layer and never upgraded by the report writer. Negative, null and uninterpretable results are preserved rather than dropped.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Abstention when the target is out of scope
Onkydra refuses and says why, rather than producing a confident answer about a target it has no grounded evidence for. A refused run does not consume the daily run ceiling.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Full artefact export
The complete run artefact, including inputs, tool calls, evidence classes and citations, exportable as JSON so a result can leave the product and be checked by somebody who does not have an account.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
Intervals resampled over the real anchor
Every interval is bootstrapped over the 60 real cases, never over the simulated draws. A simulated co-mutation profile is a Monte Carlo draw from one fitted model and is never a patient, which is also why nothing here is ever priced by how many of them a run makes.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included

The experiment

Prediction frozen before the result is known
The prediction ledger writes a claim, its evidence and its date before anybody knows the answer, so it cannot be edited to fit the outcome afterwards. The ledger is real and no prediction has been resolved yet, because no result has come back.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included
The experiment designed with us
On the subscription tiers you decide what to test and you carry the interpretation. On a programme the experiment is specified in the project plan with us before it runs, which is the difference between a tool you operate and an answer somebody stands behind.
Free
Not included
Academic lab
Not included
Commercial seat
Not included
Commercial team
Not included
Programme
Included
Observed result filed against the prediction
The result that comes back goes into the immutable record against the prediction that preceded it, whether it agrees or not. Resolution is built and has never run, because nobody has returned a result yet, and this table will say so until one does.
Free
Not included
Academic lab
Self-filed
Commercial seat
Self-filed
Commercial team
Self-filed
Programme
Filed with us, against the project plan
A named person accountable for the answer
The credence-goods literature is the honest framing here: you cannot verify this kind of answer even after delivery, and the two things that make such a market work are liability and verifiability. The subscription tiers sell capability, so you carry the risk of having used it wrongly. The programme tier sells somebody standing behind the answer. Different jobs, which is what makes the multiple coherent instead of insulting.
Free
Not included
Academic lab
Not included
Commercial seat
Not included
Commercial team
Not included
Programme
Included

Collaboration and seats

People included
The academic tier is priced per lab rather than per person because a research group of three to fifteen churns people by design, and two 10-K filings in this sector state plainly that seat pricing shrinks when a customer loses staff. Charging a lab for that is charging it for being a lab.
Free
1
Academic lab
Unlimited, in one named lab
Commercial seat
1 named user
Commercial team
3, then a flat rate for each extra
Programme
Agreed in the project plan
Shared programme memory
IN BUILD, LIVE ON NO TIER. The workspace is single-user today and there is no team or seat model behind it. It is listed because it is what the team tier is for, and it is marked because nobody should buy it expecting it this month.
Free
Not included
Academic lab
Not included
Commercial seat
Not included
Commercial team
In build, not live
Programme
In build, not live
Commercial use of the output
The academic tier is fenced on employer and IP ownership rather than on an email domain. Work inside your employment at a for-profit, and contract research where a for-profit owns the intellectual property, both need a commercial licence even when the person doing it sits in a university.
Free
Not included
Academic lab
Not included
Commercial seat
Included
Commercial team
Included
Programme
Included

Support

Where questions go
All customer-facing founder time comes out of one weekly ceiling, and support and scientific authoring draw on the same hours. The ceiling is a stated decision rather than a measurement, because there are no customers to measure. It is published so that a promise this product cannot keep is visible before you buy rather than after.
Free
Public docs and the capability register
Academic lab
Email
Commercial seat
Email, plus an onboarding walkthrough
Commercial team
Email, onboarding, and a quarterly programme review
Programme
A named contact, inside the project plan
Founder time is capped, and the cap is published
Five hours a week of customer-facing founder time, in total, across every customer. A ceiling and not a target, and it exists because the alternative failure is invisible: hours quietly leaving the research one reasonable half-hour at a time.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included

Procurement

Card, monthly, cancel anytime
Monthly billing keeps every transaction under the university purchasing-card limits observed across the institutions checked ($2,500 to $5,000) and under both stated purchase-order floors, so the purchase never has to touch procurement. Watch one clause if you are at a US university: at least one bars services from the card entirely, at any amount.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Not included
Annual billing at a discount
Seventeen percent off the monthly rate, billed yearly. It clears the academic purchase-order floor on purpose, for a buyer who would rather pass through procurement once a year than expense a card charge every month.
Free
Not included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Annual by default
Invoice and purchase-order billing
The honest upgrade gate for a small commercial buyer. It is not a feature anybody wants: it is something their organisation demands, which is exactly why single sign-on works as a gate for larger vendors and cannot work here, since a solo PI and a three-person team have no identity provider to ask.
Free
Not included
Academic lab
Not included
Commercial seat
Included
Commercial team
Included
Programme
Included
Data-processing agreement
The standard template and the sub-processor register are published and readable by anyone. A counter-signed agreement, and a customer-specific annex where one is needed, come with a commercial licence.
Free
Template published
Academic lab
Template published
Commercial seat
Counter-signed
Commercial team
Counter-signed, with an annex where needed
Programme
Counter-signed, with an annex where needed
Security questionnaire completed
We complete your questionnaire, including the HECVAT that several US universities require before any software is acquired, free items and renewals included. None has been completed yet, because no customer has asked for one.
Free
Not included
Academic lab
Not included
Commercial seat
On request
Commercial team
Included
Programme
Included
Validated-system documentation
NOT OFFERED AT ANY TIER, and it is in this table so that nobody assumes it comes with the largest one. Onkydra is a hypothesis generator and a prioritisation engine, research use only, and it makes no 21 CFR part 11 or EU Annex 11 claim. If your process needs a validated system, this is not one.
Free
Not included
Academic lab
Not included
Commercial seat
Not included
Commercial team
Not included
Programme
Not included
Research use only
Not a diagnostic, not a medical device, not a treatment recommender, and not a substitute for wet-lab validation. This does not change at any price, and a tier that relaxed it would be selling something Onkydra is not.
Free
Included
Academic lab
Included
Commercial seat
Included
Commercial team
Included
Programme
Included

Indication roadmap

Cohort packs: which indication should we build next?

Every plan ships with the DMG anchor cohort. Diffuse midline glioma is the only indication live today, and it runs within a stated scope: an n=60 anchor, 13 driver features, and no outcome layer. The rest are on the roadmap in the order below. Curating an indication means licensing and validating a whole molecular-profile library, so it is scoped as a programme rather than a monthly add-on. Tell us which one you need and it moves up the queue.

Tell us which one →

Diffuse midline glioma
Included
H3 K27M anchor, n=60, DKFZ + CPTAC
Included
Ewing sarcoma
In the queue
EWSR1 fusion, adolescent + young-adult anchor
Roadmap
MPNST
In the queue
NF1-driven, malignant peripheral nerve sheath
Roadmap
ATRT
In the queue
SMARCB1 loss, atypical teratoid rhabdoid
Roadmap
High-risk neuroblastoma
In the queue
MYCN-amplified, paediatric
Roadmap
Anaplastic thyroid
In the queue
BRAF V600E, TP53 co-mutation
Roadmap
Rhabdomyosarcoma
In the queue
PAX-FOXO1 fusion, paediatric
Roadmap
Pediatric AML (KMT2Ar)
In the queue
KMT2A-rearranged, infant + paediatric
Roadmap

A pack leaves the queue once its anchor cohort has been curated and validated end-to-end through the run pipeline. Diffuse midline glioma is the only pack live today, and it runs within a stated scope. Every other indication listed here is roadmap, not shipped, and ships in order.

Monthly by default

Billed monthly, cancel anytime, no lock-in. Annual is 17% off and is a choice rather than the default, because a monthly transaction is a small one.

Nothing is for sale yet

Checkout is not open. No card is taken anywhere on this site, and there is no plan to manage from the workspace.

Research use only

A hypothesis and prioritisation engine. Not a diagnostic, not a treatment recommender, not a medical device.

More on trials, billing, and commercial terms in the pricing FAQ.

PRIVATE BETA

Get early access to Onkydra.

Onkydra is in private beta and the workspace is open to one address. Join the waitlist and we will reach out with a first run for your programme when it opens to biotech and research teams working on rare cancers.

Plans for the teams taking on rare cancers.

One ladder, from a free benchmark to a named programme.

Every line, on every tier.

Cohort packs: which indication should we build next?

Get early access to Onkydra.

Ready to prioritise a target?

Product

Learn

Contact

Legal