LAUDOS.AI

Product changelog

Release notes are published in Portuguese, the working language of our users. The product ships continuously across REPORT (dictation to structured report), CRIT (critical findings), GUIDE (classifications), LaudAI (templates) and integrations (HL7, FHIR, DICOM-SR, worklist).

Read the changelog in Portuguese →

Recent changes

Entries are written in Portuguese, as published.

August 8, 2026 · b8ccdd1

Perguntas frequentes de soluções e artigos passaram a usar títulos semânticos, mantendo o texto e o desenho.

Clinical impact: Nenhuma mudança em regra, recomendação ou conteúdo clínico; melhora a extração e a navegação semântica.

August 8, 2026 · a06c6a5

O gerador Markdown deixou de remover tabelas numéricas de referência clínica ao filtrar arte ASCII.

Clinical impact: Restaura valores de referência em três templates de ultrassom com Doppler nos artefatos Markdown e llms-full.txt.

August 7, 2026 · 350ed7c

Conteúdo público foi simplificado e as verificações de estilo e da jornada de aquisição foram corrigidas.

Clinical impact: Sem alteração declarada de conteúdo clínico; revisão de clareza textual.

August 7, 2026 · 250477e

robots.txt ganhou proteção automatizada contra divergência entre grupos de crawlers e contra diferença do artefato publicado.

Clinical impact: Nenhum.

Four modules, one clinical flow

REPORT · PRODUCTION

Voice findings become a structured draft for physician review, editing, validation and signature.

CRIT · CONTROLLED PILOT

Detection depends on physician-provided findings; channels, timing and closure are validated per deployment.

GUIDE · BETA

Classification suggestions sourced from RadCommons and always subject to physician review.

LaudAI · BETA

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

Medium risk, not SaMD

Assistive application under CFM 2.454/2026; no autonomous image interpretation, diagnosis, signature or release.

Data flow by deployment

Processing, transmission, persistence, logs, retention, backups and regions are documented separately.

Evidence with limits

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.

52 s
MEDIAN · EDITOR TO SIGNATURE
54.6%
OF REPORTS IN UNDER 1 MINUTE
≈10 h
RETURNED PER 100 REPORTS
RadCommons LaiBench Templates Book a demo

See Laudos.AI on your own workflow

Live demo with your exam types and your template. In Portuguese, English or Spanish.

Book a demonstration

The times cited on this page come from production measurement, from opening the editor to signature. The event, the window and the N are in Metrics methodology.

Content updated on .