Find the data you're looking for
Browse open datasets and rare cohorts from hospitals, biobanks, and research institutions. Each listing shows what’s in the data, who can use it, and for what purpose.
Planning a trial? Start with a feasibility count before you license anything.
Built for teams that need rare data
- AI companies: Can we train a commercial model on this? Every listing states commercial use, model training and whether weights may leave, as yes, no or unknown. Bring your model to the data and take out the performance.
- Pharma and RWE teams: See the n per modality, treatment and response data, and length of follow-up before you request access.
- Academic researchers: Consent terms and the original source are linked on a plain-language listing you can share with your principal investigator.
For model training
What you can run, at what scale, and what you keep.
From finding a dataset to running your analysis
- Browse: Search open datasets and rare cohorts, each with a clear listing. Sign in to see the full catalog.
- Ask questions: Ask about cohort details, consent limits, or whether the data fits your project. Our team replies by email, usually within two business days.
- Request access: Request access directly from the listing. The data holder reviews every request. Approved access is licensed on a standardized agreement framework.
- Analyze in place: Bring your own container and run your analysis in your own cloud environment. Every access is logged.
Not technical?
Ask us for a count. We return aggregate numbers, and the dataset is accessed in place.
What you take out
The dataset stays with the data holder. You leave with the science: the statistics, associations, signatures and evaluation results you need to publish, file or build on.
- Tier 1 Standard outputs: Released when they are in your approved output manifest, consistent with consent, with small counts suppressed.
- Cohort counts by criteria
- GWAS and QTL summary statistics
- Effect sizes and survival results
- Correlation matrices and signatures
- Figures and aggregate reports
- Tier 2 Reviewed outputs: Released after a disclosure review, because they carry more re-identification risk.
- Model evaluation in place
- Feature-level embeddings
- Results from small or rare cohorts
- Tier 3 By exception: Only under the model weights policy, with the data holder’s written approval.
- Trained model weights
- Fine-tuned checkpoints and adapters
- Never Never released: Nothing that could rebuild or stand in for the dataset.
- Row-level or sample-level records
- Synthetic or derivative datasets
Every output comes with its context: the datasheet version, the consent codes it ran under (GA4GH DUO), the pipeline image fingerprint, the output manifest and the audit record. Drop it straight into a methods section.
What you'll see in each dataset
Example view · not a real dataset
Lung adenocarcinoma cohort
Example dataset card
Academic medical center · name shown after approval
- Disease: Non-small cell lung cancer · adenocarcinoma · Verified
- Modalities · n: WES 412 · RNA-seq 388 · H&E slides 450 · clinical 450 · Verified
- Key counts: Stage III 162 · Stage IV 288 · EGFR+ 97 · first-line immunotherapy 214 · Holder stated
- Treatment: Line of therapy, regimen and dates · Holder stated
- Response: RECIST best response · progression-free and overall survival · Holder stated
- Dates · follow-up: 2015–2024 · median follow-up 31 months · Verified
- Commercial use: Yes · Verified
- Model training: Unknown. Ask us. · Unknown
- Weights may leave: Unknown. Ask us. · Unknown
- Access model: Accessed in place · Verified
Field labels:
Verified · Holder stated · Unknown
The dataset description
See the details upfront.
See the cohort, modalities, consent scope and known limitations up front, all without exposing the underlying data.
Ask a question
Still not sure? Ask us.
Is longitudinal follow-up included in the cohort?
Yes. Median follow-up is 31 months, and the datasheet lists the date range.
Illustrative
Ask about cohort details, consent limits, or whether the data fits your project. Our team replies by email, using the listing and its documentation, so the data itself stays private.
Open dataset
Preview the catalog
These open datasets show what every card looks like. Rare cohorts are listed after you sign in, each accessed in place with the data holder's approval.
- 1000 Genomes Project: IGSR
- People: 2,504
- Modality: Genomics
- Commercial use: Yes
- Model training: Yes
- Access model: Open access
- Single-cell RNA-seq in lupus: NCBI GEO, GSE135779
- People: 58 (41 patients, 17 controls)
- Modality: Single-cell
- Commercial use: Unknown. Ask us.
- Model training: Unknown. Ask us.
- Access model: Open access
- NHANES: CDC, 1999 to present
- People: About 5,000 each year
- Modality: Population survey
- Commercial use: Unknown. Ask us.
- Model training: Unknown. Ask us.
- Access model: Open access
Find a specific dataset
If you're looking for a dataset that isn't in the catalog yet, share the data type, cohort criteria and timeline you need, and we’ll reach out to institutions that may hold it.