Visualização de leitura
October 2026 Cyber Attacks Statistics
Hugging Face’s CEO Says His P(Doom) Is ‘Close to Zero’
November 2026 Cyber Attacks Timeline
Zoomtopia 2026: Zoom Introduces AI Agents, New Security Controls
Zoom unveils AI agents and its Unified Context Layer at Zoomtopia 2026. Learn how its new automation and security tools could affect IT teams.
The post Zoomtopia 2026: Zoom Introduces AI Agents, New Security Controls appeared first on TechRepublic.
Oracle Critical Patch Update October 2026 Fixes 1,209 CVEs
Oracle Critical Patch Update for October 2026 fixes 1,209 CVEs, including six CVSS 10.0 bugs such as CVE-2026-101809 in E-Business Suite.
Related Posts:
- Atlassian Patches 176 Flaws in Jira, Confluence and Other Data Center Products
- Oracle E-Business Suite Flaw CVE-2026-101809 Scores CVSS 10.0
- Apache CouchDB 3.5.3 Fixes Six Flaws as Retired Apache Cocoon Gets Four Unpatched Bugs
The post Oracle Critical Patch Update October 2026 Fixes 1,209 CVEs appeared first on Daily CyberSecurity.
Update Chrome: Google fixes 37 security flaws, including seven critical bugs
On October 13 and 14, Google issued several Chrome updates.
Most users are on Chrome’s Stable channel. The desktop update brings Chrome to versions 155.0.8059.79/.80 for Windows and Mac and 155.0.8059.79 for Linux. The update includes 37 security fixes, including seven rated Critical.
On the same day, Google released version 155.0.8059.79 for Android. It may not be available on Google Play for everyone yet but keep an eye out for it. Google says it also includes stability and performance improvements.
Updates will roll out over the coming days and weeks.
How to update Chrome
The easiest option is to allow Chrome to update automatically. But you can end up lagging behind on updates if you never close your browser or if something goes wrong, such as an extension preventing the update.
If you don’t want to wait for the rollout to reach you, manually updating is easy. Open About Google Chrome from Chrome’s settings or Help menu. Chrome will check for updates. If one is available, click Relaunch to apply it.

You can find an explanation of the version numbering system and step-by-step instructions in our guide: How to update Chrome on every operating system.
Technical details
Some of the security fixes deserve a closer look.

Mobile protection, anywhere, anytime.
The first is tracked as CVE-2026-110152, a buffer overflow in ANGLE in Google Chrome on Android. It could allow a remote attacker to execute arbitrary code outside the sandbox via a crafted HTML page.
ANGLE is Chrome’s graphics-translation layer. A buffer overflow occurs when a program puts more data into an area of memory than it can hold, overwriting data in adjoining memory. An attacker can exploit this to make the vulnerable process run malicious code.
Chrome vulnerabilities that enable remote code execution outside the browser sandbox are particularly valuable to attackers because they can turn a visit to a malicious or compromised website into direct code running on the underlying operating system, often without requiring additional exploitation steps.
Then I spotted two use-after-free vulnerabilities in the PDF component. CVE-2026-110136, rated Critical, could allow a remote attacker to execute arbitrary code outside the sandbox via a crafted HTML page. CVE-2026-110147, rated High, could allow an attacker to execute arbitrary code inside the sandbox via a crafted HTML page.
A use-after-free vulnerability occurs when a program attempts to access a memory location after it has been freed. That can cause crashes or, in some cases, allow an attacker to run their own code.
Inside the sandbox means the malicious code would run in a restricted, sealed-off environment designed to limit access to the rest of your device. That limits what the attacker can do as it constrains them to the browser, lowering the impact compared with code running outside the sandbox. However, attackers often chain multiple vulnerabilities together to escape the sandbox and achieve a more serious compromise. So, the phrase describes an important security limitation, but it does not mean the vulnerability is harmless.
Stop threats before they can do any harm.
Malwarebytes Browser Guard blocks phishing pages and malicious sites automatically. Free, one click to install. Add it to your browser →
Q3 2026 Cyber Attacks Statistics Infographic
Q3 2026 Cyber Attacks Statistics
Update Windows: Record Patch Tuesday fixes flaw already under attack
Microsoft’s October Patch Tuesday release addresses a record-breaking 1,413 vulnerabilities, including an actively exploited Windows Shell zero-day and two other publicly disclosed flaws.
Just when you thought a previous record would stand for a while, Microsoft comes out again with the largest monthly Patch Tuesday release. Microsoft has said its growing use of AI to find and fix security vulnerabilities means users should expect more fixes in its monthly updates.
How to apply patches and check if you’re protected
These updates fix security problems and help keep your Windows PC protected. Here’s how to make sure you’re up to date:
- Click the Start button, then open Settings.
- Select Windows Update (usually at the bottom of the menu on the left).
- Click Check for updates. Windows will search for the latest security updates. If you’ve enabled Get the latest updates as soon as they’re available under More options, you may be prompted to restart immediately to complete the update.

- Otherwise, continue to the next step.
- If updates are available, they’ll start downloading automatically. When they’re ready, click Install or Restart now if prompted. Your computer may need a restart to finish the update.
- After restarting, go back to Windows Update and check again. If it says You’re up to date, you’re all set.

Technical details
As usual, we’ve selected a few vulnerabilities to highlight for readers interested in the technical details.
We’ll start with the actively exploited vulnerability tracked as CVE-2026-96951 (CVSS score 8.8 out of 10), a Windows Shell security feature bypass vulnerability.
The security feature it can bypass is the Mark of the Web (MOTW) defense, which can warn users about potentially unsafe downloaded files.
Successful exploitation requires user interaction. For example, in an email or instant message attack, the attacker could send the targeted user a specially crafted file designed to exploit the vulnerability, but the user would have to click a link or open an attachment.
Another one to keep an eye on is CVE-2026-108398 (CVSS score 8.8 out of 10), where an untrusted search path in .NET Framework allows an authorized attacker to bypass a security feature locally. The bypassed feature is Windows Defender Application Control (WDAC), which could lead to the ability to run unauthorized applications on target systems. This vulnerability was publicly disclosed before the patch.
Another publicly disclosed vulnerability is CVE-2026-92707 (CVSS score 5.3 out of 10). The flaw can let an unauthenticated remote attacker exhaust memory in affected applications that process specially crafted MessagePack data. MessagePack-CSharp is a software component that helps applications pack and unpack data in a compact format, making it efficient to store or exchange information. Exposure depends on how a service uses the library to pack and unpack the information.
From reporting threats to removing them.
Cybersecurity risks should never spread beyond a headline. Keep threats off your devices by downloading Malwarebytes today.
AI Is Now Fighting on Both Sides of the Cyber War, Singapore Warns

SICW 2026 Opens With a Warning
Fong, attending her first SICW as CSA chief, said Singapore's success has been built on being open and connected. That leaves its economy and society dependent on digital infrastructure, platforms and supply chains that extend well beyond its shores. Governments, infrastructure operators, cloud providers and cybersecurity firms each see part of the threat landscape, she said, but none sees all of it. "What appears to be an isolated incident to one organisation may in fact be part of a much larger campaign," Fong said. Assembling the picture is only half the challenge, she added. It must be done quickly enough to act before the adversary does.AI Cuts Both Ways at SICW 2026
Fong said Singapore cannot simply keep adding analysts as threats grow in volume and sophistication. Instead, it wants AI to help defenders "see more, understand faster and to act more effectively." But adversaries can harness AI too, operating at greater speed and scale. The real test of AI cybersecurity, she said, is making sure the technology strengthens defences rather than tilting the advantage towards attackers. She also flagged a newer risk. As agentic AI systems gain access to data, applications and tools, they create new dependencies and vulnerabilities. "For years, cybersecurity has required us to understand what is connected to what. In an increasingly agentic world, we must also understand what can act on what," Fong said, pointing to questions of permissions and consequences.SICW 2026 Puts AI Risks Into Practice
Rather than only discussing what AI could mean for security, CSA is testing it in practice this year. CSA is partnering with Google on an agentic AI cyber scenario exercise for international partners. It is working with the Singapore AI Safety Hub on a frontier AI governance workshop for officials from ASEAN and the wider Asia-Pacific. With Cloudflare, it is running hands-on training in AI-enabled cyber defence for critical infrastructure operators. Fong pointed to the ASEAN CERT Incident Drill, held since 2006, as proof that practising together builds the habits that matter in a real crisis.The Trust Deficit Behind SICW 2026
On 12 October, Teo described what she called a paradox. At precisely the time when more trust is needed, trust is becoming harder to sustain. AI systems can make mistakes, be manipulated or behave in unintended ways, she said, and their capabilities are advancing faster than our understanding of the risks. Teo cited UNC3886's campaign against Singapore's telecommunications sector as a recent example of sophisticated actors targeting critical infrastructure. Ransomware and cybercrime remain pervasive, while strategic competition is forcing harder choices on technologies, suppliers, standards and interoperability. "We are willing to interact more, but find ourselves willing to trust less," she said. Without trust, Teo warned, businesses will hesitate to digitalise, citizens will be wary of new technologies, and countries will be less willing to stay open and connected."Accountability Cannot Be Outsourced"
Teo said industry must be a partner in securing the digital ecosystem, since it builds and operates much of the infrastructure economies depend on. That means shared standards, threat intelligence sharing, incident reporting and testing new technologies before deployment. On increasingly autonomous AI, she said human judgement, oversight or control must be retained where necessary. "Ultimately, accountability cannot be outsourced. Least of all, to machines," she said. She outlined steps already under way. CSA released its Securing Agentic AI Addendum earlier this year to guide organisations on securing autonomous AI systems. The agency is also working with industry to prepare critical infrastructure for the transition to quantum-safe systems. In addition, leading operational technology manufacturers and technology providers have committed to attain Singapore's Cyber Trust Mark.What Comes Next
Teo said Senior Minister K Shanmugam would launch the Singapore Cybersecurity Strategy 2026, the country's updated national strategy, the following day, with partnerships remaining central to it. That approach will also carry into 2027, when Singapore assumes the ASEAN Chairmanship. Teo said Singapore will work with the bloc's 11 member states to strengthen collective cyber resilience, deepen practical cooperation and capacity building, and work more closely with industry. Estonia's Minister of Justice and Digital Affairs, Liisa-Ly Pakosta, also joined the week's proceedings for a fireside chat with Fong. Both leaders acknowledged that geopolitical competition is making cooperation harder. "Disagreement need not prevent dialogue, and competition need not preclude practical cooperation," Fong said. Teo put it simply: trust is not built through declarations, but through candid exchanges, competence and consistent action. Fong closed with a challenge to every delegate in the room: look beyond your own vantage point and find the people who can help you see the rest.Microsoft Patches More Than 1,400 Security Flaws
Microsoft today issued updates to fix more than 1,400 security vulnerabilities in its Windows operating systems and other software. Today’s epic patch batch dwarfs September’s record of nearly 1,000 fixes, which was already about ten times the normal monthly patch load from Redmond.

Just one of the weaknesses patched this week is considered a “zero day,” or already exploited in the wild. CVE-2026-96951 is a security feature bypass in the Windows Shell that affects Windows 10 and 11 systems, as well as some Server versions.
Windows 10 users might be wondering if they can still expect security updates after this month. Microsoft had been slated to stop shipping updates for Windows 10 systems in October 2025, but ended up adding another year of free updates after considerable pushback. A few months ago, Microsoft quietly extended that support for another year (until October 2027).
At least 78 of the flaws fixed this month are rated “critical,” meaning they could be exploited with little or no action by the user, and most of those critical bugs allow an attacker to execute code of their choice on the vulnerable system.
While Microsoft said there are no signs anyone is exploiting it yet, a security feature bypass in the .NET Framework (CVE-2026-108398) was publicly disclosed prior to today, potentially giving attackers advance time to work out how to exploit the bug.
Satnam Narang at Tenable said the number of flaws fixed this month far surpasses the previous record-setting year 2020, when Microsoft patched 1,261 CVEs.
“We’ve now crossed the 4,000 CVE mark with two months left to go,” he said.
Microsoft credits large language models for the high volume of bug discovery, and presumably AI gets credit for many of the patches used to fix these flaw. Meanwhile, security experts have been warning that these unusually large patch volleys from Microsoft are placing extra strain on corporate security teams, which often have to test individual updates before deploying them across a network to ensure the fixes don’t cause problems for other software.
Tyler Reguly, associate director of security research and development at Fortra, notes that just 26 Microsoft vulnerabilities appeared in the “Known Exploited Vulnerabilities” (KEV) catalog maintained by the U.S. Cybersecurity & Infrastructure Security Agency (CISA) so far this year, and security teams would do well to prioritize patching those 26 vulnerabilities first.
“Security leaders should focus on exploitable risk, and ask whether their teams have the clarity, capacity and support to act,” Reguly said.
Which vulnerabilities deserve the most scrutiny in your organization? The folks at Cisco Talos have a useful roundup of the more noteworthy flaws that Microsoft fixed today, and of course the SANS Internet Storm Center has a reliable roundup that indexes vulnerabilities by severity, CVE number and affected software.
September 2026 Cyber Attacks Statistics Infographic
September 2026 Cyber Attacks Statistics
October 2026 Cyber Attacks Timeline
Europe’s Police Chiefs Face Growing Threat of AI-Driven Crime

European Police Chiefs Convention 2026 Tackles AI-Driven Crime
A major focus of the European Police Chiefs Convention 2026 was AI in law enforcement and the changing role of artificial intelligence in criminal activity. According to Europol, criminals are using generative AI, deepfakes and autonomous agents to increase the scale and speed of their operations. These technologies are being misused in fraud, cybercrime, migrant smuggling and child exploitation. The convention also highlighted the potential benefits of AI for police agencies. Law enforcement can use the technology for faster data analysis, forensic investigations and intelligence gathering. However, participants said stronger international cooperation, investment in AI capabilities and regulatory frameworks will be needed as law enforcement agencies adopt the technology. [caption id="attachment_114531" align="aligncenter" width="600"]
Image Source: Europol[/caption]
Deepfakes and AI Expand the Criminal Threat Landscape
The growing use of deepfakes was discussed as part of a broader shift in how criminals can use AI to support illicit activities. Europol said AI is amplifying criminal capabilities, with autonomous agents and generative AI allowing operations to be carried out at greater speed. The discussions placed these developments alongside existing concerns around online fraud and cybercrime. The convention did not frame AI only as a criminal threat. Law enforcement agencies are also looking at how the same technology can improve investigations and intelligence gathering. Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security, and Democracy at the European Commission, said the origin and control of technologies used for critical law enforcement capabilities are important considerations. She also highlighted the need to understand and reduce strategic dependencies affecting security infrastructure.Lawful Access to Data Remains a Police Challenge
Another key topic was lawful access to data, alongside wider questions around information sharing and cross-border cooperation. Law enforcement representatives discussed how agencies can strengthen their ability to respond to increasingly digital criminal activity while working across jurisdictions. Wim Mijs, CEO of the European Banking Federation, also called for greater cooperation between banks, law enforcement, social media platforms and telecommunications companies to identify and stop online fraud before money is transferred. [caption id="attachment_114533" align="aligncenter" width="600"]
Image Source: Europol[/caption]
The discussions reflect the growing importance of cooperation between public authorities and private-sector organisations in responding to digitally enabled crime.
Radicalisation and Online Propaganda
The convention also examined the prevention of radicalisation, with law enforcement officials discussing online radicalisation, lone actors and changing forms of extremist activity. Europol said law enforcement faces challenges in identifying behavioural changes, dealing with encrypted communications and responding to rapidly changing propaganda. Algorithms and social isolation were also identified as factors that can amplify risks. Participants highlighted cross-border information sharing, community trust, platform accountability and early intervention as areas requiring attention.Europol and Customs Agency Strengthen Cooperation
Police and customs representatives also discussed ways to disrupt cross-border criminal activity. Europol and the World Customs Organization signed a Joint Cooperation Action Plan for 2026–27 covering strategic analysis, information exchange and capacity building. The convention concluded with the presentation of the 2026 Europol Excellence Awards in Innovation. Law enforcement agencies from Spain, Denmark and the Netherlands received awards across three categories covering ethics, diversity and inclusion, innovative operations and technical solutions. The discussions at the European Police Chiefs Convention 2026 highlighted how law enforcement agencies are adapting to a threat environment increasingly shaped by digital technologies, AI and cross-border criminal activity.OpenAI DevDay 2026: Elevating the AI Subscription Paradigm
Explore the new OpenAI Pro 500 tier unveiled at DevDay 2026. Discover the Dots AI assistant, Sign in with ChatGPT, and the efficient GPT-6.1 Sol.
Related Posts:
- Cloudflare Launches Free Certificate Authority Emphasizing MTC
- Firefox Redesign Compact Mode Returns With AI Off Switch
- OpenAI Dots AI Assistant Gets Its Own Cloud Computer
The post OpenAI DevDay 2026: Elevating the AI Subscription Paradigm appeared first on Daily CyberSecurity.
OpenAI DevDay 2026: The Shift Toward Enterprise Collaboration
Discover how OpenAI enterprise collaboration tools like ChatGPT Space and Dots are transforming workplace productivity and challenging Microsoft Copilot.
Related Posts:
- OpenAI DevDay 2026: Elevating the AI Subscription Paradigm
- Firefox Redesign Compact Mode Returns With AI Off Switch
- OpenAI Dots AI Assistant Gets Its Own Cloud Computer
The post OpenAI DevDay 2026: The Shift Toward Enterprise Collaboration appeared first on Daily CyberSecurity.
September 2026 Cyber Attacks Timeline
Transcritor ProfAMR 2026
APP: Transcritor ProfAMR 2026: a audiência criminal em texto, no seu computador GRATUITO
A dor que o Transcritor ProfAMR 2026 resolve
Um dos grandes desafios dos processos criminais é a transcrição dos depoimentos gravados nas etapas de investigação criminal e judicial [audiências]. As ferramentas de transcrição padrão exigem instalação e habilidades operacionais que estão fora do escopo da maioria dos agentes processuais. Além disso, o uso de Inteligência Artificial Generativa [IAGen] acarreta o risco de vazamento de dados, pois os depoimentos ficam sob a custódia de provedores internacionais, o que viola a Res. 615/CNJ e também a Lei Geral de Proteção de Dados.
Pensando nisso, nas pesquisas que realizo no PPCJ/UNIVALI, desenvolvi uma ferramenta de apoio à transcrição de depoimentos que roda localmente e pode ser baixada gratuitamente para o seu desktop, utilizando uma versão do Ollama em português para melhorar os resultados. No entanto, toda saída [transcrição] deve ser conferida. Use sob sua responsabilidade e supervisão. O Transcritor ProfAMR 2026 converte a gravação de uma audiência criminal em texto no computador do usuário. O áudio permanece na máquina. A versão 0.2.2 está disponível para download gratuito [free], com código aberto sob a licença MIT.
O que é?
O Transcritor ProfAMR 2026 é um programa para Windows que transcreve áudio e vídeo de audiências e depoimentos em português. O programa reúne 3 camadas: [a] um motor de reconhecimento de fala [faster-whisper, baseado nos modelos Whisper]; [b] uma camada penal de regras que corrige a grafia de termos jurídicos e padroniza a citação de dispositivos; e, por opção do usuário, a revisão por um modelo de linguagem executado pelo Ollama no computador do usuário. A ferramenta foi desenvolvida com foco em matéria criminal.
Para que serve?
O programa entrega o texto de uma audiência gravada, com o minuto de cada fala entre colchetes. O texto serve de apoio para localizar trechos, preparar voto, decisão, parecer, memorial ou petição e citar a fala com a marcação de tempo. O público-alvo é o de agentes processuais [investigadores, acusadores, defensores, peritos, julgadores]. A gravação continua a ser a fonte do ato; a transcrição encurta o caminho até o trecho procurado.
Como funciona?
O usuário escolhe o arquivo de áudio ou vídeo [mp4, mp3, wav, m4a, wma, asf, mkv etc.] e clica em Transcrever. O motor de reconhecimento de fala converte a voz em texto e marca o tempo de cada trecho. A camada penal aplica regras que o usuário lê e edita no arquivo correcoes.json: "abeas corpus" passa a "habeas corpus"; "artigo 33" passa a "art. 33"; "parágrafo 4" passa a "§ 4º"; "parágrafo 10" passa a "§ 10", conforme a regra de numeração ordinal até o nono e cardinal a partir do décimo da LC 95/1998, art. 10. As regras corrigem grafia e preservam a fala do depoente: "tava" permanece "tava".
A revisão por um modelo de linguagem corrige a pontuação e a acentuação. Uma trava recusa a proposta do modelo que altere qualquer número [artigo, data, valor, placa] ou que mude o tamanho do trecho em mais de 25%; nesse caso, o texto da camada penal permanece. O resultado sai em três arquivos: a ata em texto, com cabeçalho que registra o arquivo de origem, o hash SHA-256, o motor, o modelo e a data; a legenda em formato SRT; os dados em JSON, com o texto bruto do reconhecimento ao lado do texto final de cada trecho.
O motor de CPU identifica a fala e o tempo; a atribuição de quem fala cabe ao usuário, com base na gravação. A primeira transcrição exige acesso à internet para baixar o modelo [cerca de 484 MB no modelo padrão]. Depois, o programa opera sem conexão, diretamente no seu desktop.
Como aplicar na prática do Direito Penal e Processo Penal
O CPP, art. 405, § 2º, dispensa a transcrição do depoimento registrado por meio audiovisual e determina o envio de cópia do registro original às partes. A Resolução CNJ 105/2010, art. 2º, segue a mesma linha e faculta ao magistrado determinar a degravação. O Transcritor opera justamente aqui: a transcrição é um instrumento de trabalho de quem lê o processo; a gravação permanece como registro do ato.
Três controles orientam o uso: [a] a citação em peça traz o minuto da gravação, para que a parte contrária e o órgão julgador confiram a fala na fonte; [b] o hash SHA-256 no cabeçalho vincula o texto ao arquivo exato que o gerou; a divergência de hash aponta arquivo diferente; e [c] o processamento no computador do usuário mantém depoimento, dado pessoal e dado sensível fora de servidor de terceiro, em linha com a Lei 13.709/2018 [LGPD]. O programa oferece apoio à leitura da prova; a prova é a gravação-fonte da transcrição.
Linguagem Simples
O Transcritor funciona como um datilógrafo que escuta a gravação da audiência e escreve o que ouve, com a marcação da hora de cada fala. Depois, um revisor de termos jurídicos corrige a grafia de termos e expressões, como "habeas corpus" e "art. 33". Tudo acontece dentro do computador. O texto ajuda a encontrar a fala; antes de citar, a pessoa ouve o trecho na gravação.
Representação Visual do Fluxo

Captura de Tela

Janela do Transcritor ProfAMR 2026, versão 0.2.2, com a ata de demonstração. Fonte: repositório do projeto no GitHub.
Onde Baixar de Graça
Programa para Windows [versão atual]: TranscritorProfAMR-Windows.zip
Página da versão, com notas de lançamento: github.com/ProfAMR404/Transcritor-ProfAMR-2026/releases
Código-fonte e documentação: github.com/ProfAMR404/Transcritor-ProfAMR-2026
Ollama portátil [revisão por modelo de linguagem; o Transcritor baixa sozinho pelo botão Preparar Ollama]: ollama-windows-amd64.zip
Ollama, instalador para Windows: OllamaSetup.exe
POP: Procedimento Operacional Padrão
Nível 1: Operação Básica
Baixa aqui o arquivo TranscritorProfAMR-Windows.zip pelo link acima.
Clicar com o botão direito no arquivo e selecionar "Extrair tudo". Extrair a pasta inteira.
Abrir a pasta e dar um duplo clique em TranscritorProfAMR.exe. Se o Windows exibir "O Windows protegeu o computador", clicar em "Mais informações" e, em seguida, em "Executar assim mesmo".
Ler o aviso de uso e clicar em OK. O atalho na Área de Trabalho aparece na primeira abertura.
Clicar em "Escolher arquivo" e selecionar a gravação.
Manter o modelo "small" [padrão] ou escolher "base" [triagem] ou "medium" [precisão].
Clicar em "Transcrever" e acompanhar o progresso na barra inferior.
Clicar em "Copiar para o documento" e colar no editor com Ctrl+V. Para copiar um trecho, selecione-o antes.
Para a revisão por modelo de linguagem: clicar em "Preparar Ollama" e responder "Sim" às perguntas. O programa usa o Ollama instalado ou baixa a versão portátil e o modelo gemma3:4b, uma única vez.
Consulte o botão "Instruções ao usuário" para o passo a passo completo.
Nível 2: Uso no processo penal [controles de prudência]
Transcrever a partir da cópia do registro audiovisual juntada aos autos, sem edição prévia do arquivo.
Guardar a ata em texto com o cabeçalho: o hash SHA-256 identifica o arquivo que gerou a transcrição.
Localizar o trecho na transcrição e ouvi-lo na gravação antes de citá-lo.
Indicar o falante com base na gravação; o motor de CPU registra a fala e o tempo.
Conferir com atenção nomes próprios, números, datas, valores e placas.
Citar a fala com o minuto entre colchetes, para permitir a conferência pela parte contrária e pelo órgão julgador.
Tratar a transcrição como material de trabalho; a prova é a gravação.
FAQ: Perguntas Frequentes
Quanto custa?
O programa é distribuído sem custo e é de código aberto sob a licença MIT.
O áudio vai para a internet?
O processamento ocorre no computador do usuário. A web serve para baixar o programa e os modelos, uma única vez.
A transcrição substitui a gravação?
A gravação permanece como fonte do ato. O art. 405, § 2º, do CPP, dispensa a transcrição do registro audiovisual. A transcrição apoia a leitura e a citação em peças ou na gestão do caso.
O programa identifica quem fala?
O motor de CPU registra a fala e o tempo. A indicação do falante cabe ao usuário, com base na gravação.
Preciso do Ollama?
A revisão por modelo de linguagem é uma opção do usuário. A camada penal funciona sem o Ollama.
Qual computador preciso?
O projeto recomenda Windows 10 ou 11, de 64 bits, com 8 GB de memória RAM; para o modelo de revisão gemma3:12b, 16 GB.
Glossário de Termos
Camada penal: conjunto de regras fixas do Transcritor que corrige a grafia de termos jurídicos e padroniza a citação de dispositivos.
Faster-whisper: biblioteca que executa os modelos Whisper de reconhecimento de fala na CPU ou na GPU.
Hash SHA-256: sequência de 64 caracteres calculada a partir do conteúdo de um arquivo; qualquer alteração no arquivo produz uma sequência diferente.
Modelo de linguagem: programa treinado em texto que propõe correções de pontuação e de grafia.
Ollama: programa que executa modelos de linguagem no computador do usuário.
Reconhecimento de fala: técnica que converte voz gravada em texto com marcação temporal.
SRT: formato de legenda com o tempo de início e de fim de cada trecho.
Trava: verificação do Transcritor que recusa a proposta do modelo de linguagem quando ela altera o número ou o tamanho do trecho.
Antes de colar a fala no voto, ouça o minuto indicado entre colchetes. Se o texto e a gravação divergem, prevalece a gravação.
Mais uma Coisa
Confira a página www.profamr.org, onde disponibilizo conteúdo gratuito. Aliás, confira o Monitor Penal, onde todas as notícias nacionais e internacionais sobre direito penal e processo penal são reunidas, evitando a necessidade de acessar diretamente todos os sites para se manter informado. Basta dar um clique se interessar.
Referências Bibliográficas
BRASIL. Decreto-Lei nº 3.689, de 3 de outubro de 1941. Código de Processo Penal. Brasília, DF: Presidência da República. Disponível em: https://www.planalto.gov.br/ccivil_03/decreto-lei/del3689compilado.htm. Acesso em: 27 set. 2026.
BRASIL. Lei Complementar nº 95, de 26 de fevereiro de 1998. Dispõe sobre a elaboração, a redação, a alteração e a consolidação das leis. Brasília, DF: Presidência da República. Disponível em: https://www.planalto.gov.br/ccivil_03/leis/lcp/lcp95.htm. Acesso em: 27 set. 2026.
BRASIL. Lei nº 13.709, de 14 de agosto de 2018. Lei Geral de Proteção de Dados Pessoais [LGPD]. Brasília, DF: Presidência da República. Disponível em: https://www.planalto.gov.br/ccivil_03/_ato2015-2018/2018/lei/l13709.htm. Acesso em: 27 set. 2026.
CONSELHO NACIONAL DE JUSTIÇA. Resolução nº 105, de 6 de abril de 2010. Dispõe sobre a documentação dos depoimentos por meio do sistema audiovisual. Brasília, DF: CNJ. Disponível em: https://atos.cnj.jus.br/atos/detalhar/166. Acesso em: 27 set. 2026.
OLLAMA. Ollama: releases. GitHub, [S. l.], 2026. Disponível em: https://github.com/ollama/ollama/releases. Acesso em: 27 set. 2026.
SYSTRAN. faster-whisper. GitHub, [S. l.], 2026. Disponível em: https://github.com/SYSTRAN/faster-whisper. Acesso em: 27 set. 2026.