Your data is yours
By default, we don’t use your inputs, outputs, or uploaded files to train general-purpose models, and we don’t pool or share data across customers or regions. Set out in our Terms of Service.
Products
Solutions
Pricing
Resources
About
Understand how we protect your data, govern AI use and keep improving. Explore the answers, evidence and updates behind our commitments.
Your patents, your strategy, your competitive intelligence — protected by independently audited controls and strict data isolation.
By default, we don’t use your inputs, outputs, or uploaded files to train general-purpose models, and we don’t pool or share data across customers or regions. Set out in our Terms of Service.
Your data is stored and processed in a dedicated regional environment — EU/EEA or US — kept isolated and routed by region on default deployment.
Our information security is managed under a formal management system, certified to ISO/IEC 27001:2022 by SGS (UKAS-accredited) and independently audited.
Aligned with the EU AI Act: clear AI-interaction notices, machine-readable marking of AI output where required, and human review before relying on AI for high-stakes decisions.
Our Data Processing Agreement and Terms of Service set binding terms on data protection, access, and AI use — auditable and built to meet GDPR, ISO, and SOC 2 standards.
Our controls are validated by independent third parties — ISO 27001 by SGS (UKAS-accredited), ISO 37301 by BSI, and SOC 2 against AICPA criteria — not self-assessed.
“Data security is never finished, and we continue to invest in it.” At Patsnap, protecting your data is central to how we operate and to the trust you place in us.
Jeffrey Tiong
Founder & CEO, Patsnap
The latest updates and improvements to Patsnap's security and compliance.
In October 2026, we moved application monitoring for our US and EU environments to Datadog.
We are now completing the migration of agent sandbox execution environments for international users to Daytona’s US and EU regions. The target launch date is 7 November 2026, subject to final validation.
During the week of 2 November 2026, we moved the OpenAI, Claude and Gemini model deployments used by our EU data-center environments to the EU regions supported by those services.
This update concerns the model deployments used by those EU environments. The applicable processing scope and commitments depend on the service, provider and configuration.
Our Trust Center is an ongoing product: a place for clear answers, security documentation and updates as our services and controls evolve.
We will continue to collect customer questions, expand the technical Q&A, and publish relevant security news, press releases and improvements as information becomes available. We are also developing Japanese-language content so more customers can explore these resources.
Our SOC 2 Type II assessment is in progress.
Our SOC 2 Type I report remains available on request through the document portal, subject to the applicable access and confidentiality requirements.
Customer data for users in the European Union is stored on AWS in Frankfurt, Germany. For users in all other regions, customer data is stored on AWS in Virginia, United States.
Hosting location and AI processing location are different. Some AI services may process data outside your hosting region, subject to the applicable contractual protections.
Read the AI processing Q&A
Data security, privacy and AI governance.
How requests are reviewed, handled and disclosed.
Information security management system certification.
Control design assessed at a specific date.
The framework behind AI safety governance.
Principles and policies for responsible AI use.
We treat any data submitted by or on behalf of your company or its authorized users as confidential customer information. This includes prompts, queries, searches (including AI searches), conversations, uploads, documents, files, instructions, and comments.
Customer data is encrypted when stored using AES-256-based controls and protected during transmission using TLS 1.2 or higher. Access checks for each customer account, role-based permissions and network boundaries restrict who can access the data. Privileged accounts use multifactor authentication. AWS provides the underlying cloud security capabilities, and Patsnap configures and operates the protections within that environment.
An administrator manages the technical operation of our customer databases and AWS accounts, including access permissions. For the US and Frankfurt customer environments described here, the administrator is based in Singapore, our corporate headquarters, and access is controlled from there. Patsnap personnel based outside Singapore do not have access to confidential customer data held in these environments or authority to grant access to others.
Our internal policies prohibit unauthorized access to confidential customer data, and technical controls enforce these restrictions. Any exceptional human access requires prior written authorization from Patsnap and must also be authorized by the customer or required by an applicable binding legal obligation. Such access is limited in scope and duration and is logged.
For the US and Frankfurt customer environments described here, data is stored in AWS Virginia in the US or AWS Frankfurt in Germany, depending on the customer's location.
Our default web application runs as a separate instance in each location, with an isolated, encrypted customer data environment. The US and Frankfurt environments do not share customer databases or exchange confidential customer content with each other. AI processing locations may differ from where data is stored.
Developing software does not give anyone permission to view customer information or change access rights. Changes to applications and infrastructure follow documented procedures, including code review, security testing, approval and controlled deployment by authorized personnel.
Technical controls prevent Patsnap personnel based outside Singapore from using application services, Patsnap's own AI model processing, gateways, debugging logs or support systems to access confidential customer content or grant themselves additional permissions. When applications process readable customer content, software change controls and service permissions apply alongside encryption of stored data.
Patsnap does not use customer inputs, outputs or related usage data to train its models. Approved cloud AI providers are contractually committed not to use your inputs or outputs for model training.
Our primary AI processing uses Anthropic’s Claude through Amazon Bedrock within AWS. Prompts and outputs processed through Bedrock are not shared with Anthropic. Selected workflows also use OpenAI models hosted in Microsoft Azure and Google Gemini. These services may process data outside your hosting region, subject to contractual data protection and no-training commitments. Please see our subprocessor list for further information.
Your AI outputs and conversation history are stored encrypted in your hosting region. Patsnap personnel cannot access this content except through the limited, authorized access arrangements described above, which require appropriate authorization, are limited in scope and duration, and are logged.
Patsnap holds ISO/IEC 27001:2022 certification and a SOC 2 Type I report. Our security programme includes independent penetration testing, AI red teaming (testing AI systems for security weaknesses), vulnerability management, monitoring, and incident response.
Testing addresses matters such as authorization, tenant boundaries, privilege escalation, prompt injection and data exposure, and findings are managed through remediation and retesting. You can view or request the certificates and reports through our Trust Center.
Yes. AWS provides our primary hosting infrastructure and services for encryption, identity and network controls, logging and monitoring. We also use specialist providers for operational monitoring, compliance management, and independent security testing.
Patsnap remains responsible for its access policies and configuration. The information processed by monitoring and compliance tools depends on their function and configuration and may include confidential customer information. Providers that process customer data are subject to our supplier review process and applicable data processing terms.
As of November 2026, Patsnap has not received a request from any government authority for any data or information and has not provided any information in response to such a request.
Our Legal team would review any such request with our Privacy and Security teams to assess the requesting authority, the relevant jurisdiction, and whether the request is legally binding. Such a request for data would not itself grant access to our systems. We would reject a request without an applicable binding legal basis and limit any legally required disclosure to the information we are legally required to provide.
Patsnap personnel based outside Singapore have no access to confidential customer data held in the US or Frankfurt environments and therefore have no authority to grant access.
Yes, where legally permitted, we would promptly notify the affected customer of such a request relating to any information, unless a legally binding restriction prevents us from doing so. We would also follow the notification requirements in your applicable agreement and data processing terms.
Your account team can arrange a security review with Patsnap's Security and Legal teams in accordance with your agreement and the agreed review process. Relevant assurance documents can be made available subject to applicable confidentiality requirements.
A review can cover our ISO/IEC 27001 certificate, SOC 2 report, relevant independent penetration-test and AI assessment materials, the current subprocessor register, applicable transfer impact assessments, and the hosting and AI configuration for your company's services.
Contact your account team or email compliance@patsnap.com or dpr@patsnap.com.
Where required, AI-interaction notices and machine-readable marking identify AI-generated output. Applicable product terms address human review before relying on AI output for legal, patent, regulatory or other material decisions.
The requirements depend on the feature and intended use. Review the applicable service terms and ask your account team about your configuration.
Retention and deletion are governed by the applicable agreements and service policies. Requirements can differ for active content, backups, operational logs and records retained for legal purposes.
Contact your account team or dpr@patsnap.com to discuss the arrangements and request process for your service. The Privacy Policy provides further information on applicable personal information rights.
Ask about Patsnap products, pricing, implementation, security or anything else. Our highly trained reps are standing by, ready to help.
Contact us