Built by physicians,
for the radiology routine
Laudos.AI is a Brazilian healthtech founded by physicians. We build the post-imaging layer of radiology: everything that happens between the acquired image and the signed report, dictation, structuring, review, classification support and critical-finding communication.
Our founding conviction is that AI in medicine must be assistive. The system proposes structure, terminology and a draft; the radiologist reviews, edits and signs. That design is not a limitation, it is the product: it is what makes the gain in speed compatible with clinical responsibility and with Brazilian regulation (CFM Resolution 2.454/2026).
We publish evidence with its limits. RadCommons is a versioned corpus; LaiBench v3.10.0 is an internal controlled fidelity benchmark with 120 cases, not external clinical validation. Median editor-to-signature time is 52 seconds in a fixed April 9 to May 9, 2026 snapshot of 5,200 reports.
The company is based in São Paulo, Brazil and has a designated DPO. Data regions and subprocessors depend on the contracted feature and are documented for each deployment.
Four modules, one clinical flow
Voice findings become a structured draft for physician review, editing, validation and signature.
Detection depends on physician-provided findings; channels, timing and closure are validated per deployment.
Classification suggestions sourced from RadCommons and always subject to physician review.
Draft structure per modality, subject to configuration, contract and physician review.
Plugs into your existing flow
API, HL7/DICOM and Agent are different paths. Compatibility depends on system, version and connector and is validated per deployment; there is no universal integration claim.
Governance first
Assistive application under CFM 2.454/2026; no autonomous image interpretation, diagnosis, signature or release.
Processing, transmission, persistence, logs, retention, backups and regions are documented separately.
RadCommons is the cited corpus; LaiBench is an internal controlled fidelity benchmark, not external clinical validation.
Historical production snapshot: April 9 to May 9, 2026; n = 5,200. It is not a live rolling window.
Os tempos citados nesta página vêm de medição em produção, do editor aberto à assinatura. O evento, a janela e o N estão em Metodologia das métricas.
Conteúdo atualizado em .