What is JADBio?
JADBio (Just Add Data Bio) is an automated machine learning service from JADBio – Gnosis DA S.A., a company based in Heraklion, Crete, with a US office in Los Angeles. It is built for the kind of data that biologists and clinical researchers have: few samples, many measurements. You upload a table, such as gene expression, proteomics, metabolomics or clinical records, pick the outcome you care about, and it handles the rest: preprocessing, feature selection, trying different algorithms, tuning them and estimating how well the final model will do on new patients.
What comes back is a model plus the short list of features it relies on, with charts and a report you can share by link. Its support pages say it takes 15 to 200,000 samples and up to 1 million variables, and its 2022 paper in npj Precision Oncology describes how it picks small signatures while keeping honest out-of-sample estimates. Paid plans add survival (time-to-event) analysis, a REST API with a Python client, batch runs and unlimited model export.
A free Basic plan covers small projects. The Team plan is $2,199 per team a month, billed annually, for five or more people, and larger tiers can run as an AWS container or on your own servers. JADBio is also listed on AWS Marketplace.
How much does JADBio cost?
| Plan | Price | What you get |
|---|---|---|
| Basic | Free | 1 seat, 3 projects, 50 MB uploads, 500 MB storage, 1 CPU, 1 model export |
| Team 5+ | $2,199 per team/month, billed annually | Five or more seats, survival analysis, API, unlimited model export, premium support |
| Business Pro 40+ | Custom | AWS container or on-premise delivery, up to 1,000 seats |
Basic free (1 seat, 3 projects, 50 MB uploads, 500 MB storage, 1 CPU, 1 model export); 14-day Team trial; Team $2,199 per team/month billed annually (5+ seats); Team 1+ and Business Pro (AWS container or on-premise) on quote; academic and classroom plans on request.
Is JADBio worth it?
You have 120 tumour samples and 20,000 genes, and a question: which handful of them predicts who responds to treatment? A general data tool will happily give you a model that looks perfect on your data and falls apart on the next patient. With so few samples and so many measurements, that trap is the default.
JADBio is built around that problem. You upload the table, say which column is the outcome, and it tries preprocessing steps, feature selection methods and algorithms (ridge logistic regression, decision trees, support vector machines and random forests, according to its FAQ), tunes them, and reports how the winner should perform on unseen data. Its peer-reviewed paper, published in npj Precision Oncology in 2022 by JADBio's own researchers, says it typically finds signatures of just a few features while keeping its performance estimates accurate.
The output is meant to be read, not just run. You get the selected features, an interpretation of each one, model visualisations and a report a colleague can open from a link without an account. Paid plans add survival analysis for time-to-event questions such as relapse, a REST API and a Python client for scripting, and batch analyses.
The catch: the price jumps from nothing to $2,199 a month for a team, billed a year at a time, with no published monthly or single-seat price. The free Basic plan is tight (3 projects, 50 MB files, one CPU) and, according to the FAQ, cannot run survival analysis once the 14-day trial ends. Its algorithm list is classical; the FAQ says deep learning is not on offer. And the privacy policy says data processing results count as non-personal information that may be shared with other parties to improve services and for marketing, which a hospital's data protection officer will want to read closely.
Our take: think of JADBio as a careful statistician for one specific job, finding a small biomarker panel in omics or clinical data. For a lab doing that every month it can save weeks of scripting; for occasional tabular prediction, a notebook or a general AI analyst is far cheaper.
Where it earns its place
A cancer research group at a university hospital
Find a small gene panel that predicts response to immunotherapy
Upload the expression matrix with a response column and let JADBio search feature selectors and models; it returns a short signature with estimated out-of-sample performance and a shareable report. Ask about the academic discount, since Team list price is $2,199 a month billed annually.
A 20-person diagnostics start-up building a blood test
Turn proteomics data into a candidate classifier and a relapse model for investors
Use the 14-day Team trial for classification and survival analysis, then script repeat runs through the REST API or Python client. A subscription is needed to keep survival analysis after the trial.
A lecturer teaching a biostatistics module
Let 30 students build predictive models without writing code
Request the Classroom plan, which the pricing page lists with 30 floating licences and educator control over student accounts; students upload course datasets and compare the reports.
Before you commit
- Setup
- Sign up for the free Basic plan, upload a table with one row per sample, choose the outcome column and start an analysis; large or multi-class datasets can take a long time, and no run times are published.
- Integrations
- REST API and the jadbio Python client on PyPI for paid plans; connections to public repositories on Team and private ones on Business Pro; listed on AWS Marketplace.
- Data and privacy
- Hosted on AWS (region not named), encrypted in transit and at rest, kept as long as the account is active; deletion requests are answered within one month. The privacy policy treats processing results as non-personal information that may go to other parties for improving services and marketing. No ISO or HIPAA certification is published.
- Cost at scale
- Team is $2,199 per team a month, billed annually, which is about $26,388 a year for five or more seats; Team 1+ and Business Pro (up to 1,000 seats, AWS container or on-premise) are quoted.
- Cancelling and refunds
- The pricing page says you can quit anytime; Team is billed annually, and no refund terms or cancellation steps are published.
- Limits
- 15 to 200,000 samples and up to 1 million variables per dataset; Basic allows 3 projects, 50 MB uploads, 500 MB storage and one analysis at a time; image, signal and single-cell analysis are listed only on Business Pro.
What’s good
- Built for few samples and many variables, up to 1 million
- Picks small feature signatures with out-of-sample performance estimates
- Survival analysis, REST API and Python client on paid plans
- Shareable reports that open without an account
- AWS container or on-premise delivery on larger plans
What’s not
- Team plan is $2,199 a month, billed annually
- Free plan limited to 3 projects, 50 MB and one CPU
- No deep learning, by its own FAQ
- Processing results may be shared with third parties under the privacy policy
- No published security certifications or refund terms
Best for
Translational research and biotech teams who repeatedly need validated biomarker signatures from omics or clinical data and have no spare machine learning engineer.
JADBio alternatives
JADBio: common questions
- Is JADBio worth it?
- JADBio does one hard thing well, turning a small, wide biomedical dataset into a short, honestly validated biomarker list, but past the free tier it costs like lab equipment. We rate it 3.7 out of 5.
- How much does JADBio cost?
- Basic free (1 seat, 3 projects, 50 MB uploads, 500 MB storage, 1 CPU, 1 model export); 14-day Team trial; Team $2,199 per team/month billed annually (5+ seats); Team 1+ and Business Pro (AWS container or on-premise) on quote; academic and classroom plans on request.
- What are the downsides of JADBio?
- Team plan is $2,199 a month, billed annually; free plan limited to 3 projects, 50 MB and one CPU; no deep learning, by its own FAQ; processing results may be shared with third parties under the privacy policy; no published security certifications or refund terms.
- Who is JADBio best for?
- Translational research and biotech teams who repeatedly need validated biomarker signatures from omics or clinical data and have no spare machine learning engineer.
- How much does JADBio cost for a team or in production?
- Team is $2,199 per team a month, billed annually, which is about $26,388 a year for five or more seats; Team 1+ and Business Pro (up to 1,000 seats, AWS container or on-premise) are quoted.
- Is JADBio safe for company data?
- Hosted on AWS (region not named), encrypted in transit and at rest, kept as long as the account is active; deletion requests are answered within one month. The privacy policy treats processing results as non-personal information that may go to other parties for improving services and marketing. No ISO or HIPAA certification is published.
- How do I cancel JADBio, and can I get a refund?
- The pricing page says you can quit anytime; Team is billed annually, and no refund terms or cancellation steps are published.
- What are the best JADBio alternatives?
- The closest alternatives we have reviewed are Julius AI and Hex.
Answers checked Nov 20, 2026.