This page describes how SynthGraph is designed. It does not present certifications, audits or compliance claims, because there are none to present.
What this page does not claim
SynthGraph is designed with privacy and security requirements in mind. Formal legal compliance claims — SOC 2, ISO 27001, GDPR conformance and similar — require appropriate legal and audit review, and none are claimed here. If your institution needs a specific assurance, tell us in a demo request and we will tell you plainly where things actually stand.
Security principles
Most of what makes an experiment reproducible is metadata: parameters, versions, seeds, references. SynthGraph is built around storing that, which means the system does not need custody of the datasets themselves in order to be useful.
Authentication
The SDK authenticates with an API key supplied by the environment. Keys are treated as secrets: not committed to research code, not printed in output, and not written into logs.
Authorization
Authorization is enforced by the API rather than by the client. A client that asks for a record it should not have does not receive it, regardless of what the client believes about itself.
Data isolation
Ownership runs User → Project → Experiment, and access follows that structure. Provenance from one group’s project is not visible to another’s.
Credential protection
Structured logging is used for operational visibility, and credential material is excluded from it. Log output is intended to be safe to read and safe to keep.
Transport security
All API traffic is served over HTTPS. The SDK’s endpoint is configurable so the same code can talk to a hosted deployment or one inside your own network.
Logging
Logs exist to answer operational questions — whether a request succeeded, how long it took, what failed. They are not a second copy of your research data.
Privacy by design
Because the product records references and metadata rather than datasets, the smallest useful footprint is also the normal one. Nothing has to be turned on to avoid collecting more than necessary.
Storage choices
Generated datasets can remain on the researcher’s filesystem, on lab storage, or in object storage. SynthGraph records the reference, the version and — where available — the checksum.
Deployment model
The same provenance model is intended to run in a hosted deployment or one operated inside an institution’s own environment, with the SDK pointed at a different endpoint.
Deployment
SynthGraph is designed to support cloud-hosted and local/self-controlled deployments while preserving the same core provenance concepts.
SynthGraph SDK
one client, one configurable endpoint
A hosted API, so a lab can start recording provenance without operating a service.
The same provenance model, deployed where the institution’s requirements say it has to live.
Deployment targets are described in terms of where the service runs, not in terms of a specific cloud provider. That choice is not finalised.
SynthGraph is currently available through a limited research pilot. Request a demo to see how provenance fits into your existing workflow.
SynthGraph is being developed with a small number of research workflows before broader access.