Know what you’re evaluating.

Yes. Product use has no subscription fee during early access, and no credit card is required to request access. There is no announced end date. We’ll communicate future pricing in advance. Read the full offer and scope.

You can try the synthetic demo immediately, with no sign-in. To evaluate the workspace with your team, request free access. We review your workflow and confirm onboarding rather than promise instant provisioning.

The report QA workflow surfaces potential issues such as inconsistent laterality, discrepancies between report sections and missing comparison context. Findings include source evidence for review. Available checks depend on the report and configuration; a clean QA result does not establish clinical correctness.

No. Med-AI Clinical supports report review. A clinician remains responsible for checking source material, correcting the report and signing the final version. The 3D Atlas shows reference anatomy, not a patient-specific reconstruction or the extent of disease.

Radiology report QA is the main early-access workflow. The public demo also includes a synthetic laboratory case. During onboarding, agree the modalities, report language, file formats and volume your team needs; demo coverage is not a guarantee of support for every clinical report.

You can explore the demo without an integration. A workspace can receive patient report uploads and pasted report text. PACS/RIS/LIS/EHR connections require checking your vendor, interface, credentials and deployment. FHIR-oriented interfaces do not imply compatibility with every EHR. Discuss your system.

Review the source evidence, edit the report and submit feedback inside the report workspace. Clinicians can rate whether a QA alert was useful. A suggestion is not applied as a final clinical decision; the reviewer controls sign-off.

Uploaded source files use the configured S3 bucket. Report text, workflow state and related metadata are also stored in the application database. Workspace tenancy and role checks restrict access. The actual storage region, backups, provider settings and access arrangements need to be confirmed for your deployment. Read the data-flow overview.

The current application sends selected content to the configured model provider for inference and can retrieve workspace reference documents. That is different from training a model. The application also has administrator-controlled export tools for creating training datasets from reviewed reports. Export is a separate action; it does not itself train a model. Agree whether these tools may be used and how exported data is handled before onboarding. Provider retention and training terms must be verified for the chosen provider and account before patient data is processed.

There is no single published retention period covering all early-access deployments. Confirm retention and deletion procedures for the database, S3 objects, audit records, backups and model provider before uploading patient information. Contact the team to arrange a data-handling review.

We are not publishing validated accuracy, time-saving percentages, customer outcome claims or independent security certifications here. Early access is for supervised evaluation. See the measurement plan and its limitations.

Workspace administrators can upload reference documents such as protocols, guardrails and studies to the Knowledge Base. Relevant retrieved content can inform supported AI workflows. It does not replace clinician judgment or guarantee that every document will be retrieved for every report.

Review useful-alert ratings, report edits, creation-to-signature time and active participants in the workspace results dashboard. Define a baseline and inspect missed issues separately before drawing conclusions. Read how these measures should be interpreted.