Legal

Privacy

Privacy Policy for MedSafeAI

Last updated: 28 October 2026


Introduction

MedSafeAI is provided by SAFE AI NETWORK PTE. LTD. (“we,” “our,” or “us”). This Privacy Policy explains how information is handled when you use MedSafeAI’s mobile applications, website, and related services.

MedSafeAI brings medical AI models, specialty-focused assistants, and working tools into one workspace for doctors and licensed healthcare professionals. It provides information, research, and educational support. It does not replace professional clinical judgment or provide diagnosis, treatment, or emergency care. Features and data flows vary by platform, app version, selected model, and subscription.


Information We Process

  1. Account and device information: Depending on how you use the service, we process account identifiers, device identifiers, language preferences, device and operating-system information, and authentication information. If you choose Apple or Google sign-in, this may include the account identifier and email address made available by that provider. Mobile guest access can still involve a device-linked service account; it does not mean anonymous or entirely offline use. Access to the web AI workspace requires sign-in; public website pages can be viewed without an account.
  2. Content you provide: Questions, prompts, messages, clinical context, images, medical images, PDFs, other supported files, case notes, library documents, reference materials, and support requests. Filenames, document contents, and image metadata may also contain personal or sensitive information.
  3. Conversation and workspace information: Conversation history, AI responses, generated content, case and file associations, and selected model or assistant. Some ordinary chat requests also include profile answers, category context, and city or country information.
  4. Subscription information: Product and subscription identifiers, purchase and entitlement status, transaction-related information, and purchase attribution. Payment handling depends on whether you purchase through an app store or the web.
  5. Usage, diagnostics, and website information: Feature interactions, subscription events, technical errors, device and network information, and website campaign attribution. Diagnostic reports can include account or device identifiers and limited request or response details; they should not be assumed to be anonymous.

The information processed depends on the feature you use, your settings and permissions, and the services enabled for that platform. Not every category is collected from every user.


Why We Process Information

We process information to authenticate access, deliver requested AI responses and tools, maintain conversations and workspaces, manage subscriptions and entitlements, provide support and notifications, diagnose technical problems, understand service usage and acquisition, protect the service, and meet applicable legal obligations. Processing for an individual feature follows the distinctions described below.


AI Models, Assistants, and Research

MedSafeAI uses official APIs from providers including OpenAI, Anthropic, Google, and xAI to offer model families such as GPT, Claude, Gemini, and Grok. The available selection can change. Using these APIs does not imply that the providers endorse MedSafeAI or that MedSafeAI represents them.

Ordinary AI chat sends your message, selected attachments, and relevant conversation or feature context through our backend for processing. Ask All AI can send the same question and relevant content to multiple models and providers. A 360° AI Summary involves further processing of model responses. Second AI Opinion and model switching can involve another provider. Routing depends on the selected models and workflow.

Specialty-focused assistants and Evidence & Guideline Lab support medical information and research. Research requests may involve external search and source services. A citation, guideline reference, or agreement between models is not a guarantee of accuracy. You must independently verify outputs and sources.

In the mobile app, we ask for your explicit permission before sharing your content and related personal information with third-party AI providers. This may include your messages, relevant chat history, submitted images or documents, and your preferences, including your specialty and onboarding answers, as well as city or country information when used to personalize responses.

Your permission covers the current request and future AI requests involving the data and providers disclosed to you. If you decline, the request is not sent to an AI provider. You can withdraw permission in Settings to stop future sharing. Withdrawal does not automatically delete information previously shared; the retention and deletion provisions below continue to apply.


Case Workspace and Your Own AI Library

These features are not solely on-device archives. They use server-based functions to upload, list, organise, associate, rename, and delete content. Case Workspace can connect chats, files, notes, and research, and relevant case context can be used in later work. Your Own AI Library lets you upload reference documents and ask questions about them.

Saved conversations, files, and workspace information do not necessarily disappear when you close the app or end a session. Deleting an item from the interface is not a guarantee that every related log, backup, or third-party copy is erased at the same moment.


Expert Imaging and Clinical Scores and Equations

Expert Imaging includes tools for supported CT, MRI, PET, and X-ray workflows, including DICOM viewing, measurements, and reconstruction. “On-device” describes relevant local viewing, measurement, and reconstruction functions. If you request AI observations, selected rendered images, an optional report, and related context can be sent through the chat service. It does not mean that all imaging AI runs locally or that images can never leave your device.

Clinical Scores and Equations performs supported OCR and calculations on the device. You must review extracted and entered values. If you select “Ask AI to explain,” the entered values, result, formula, and source context can be sent through the AI chat service. The original document is not attached by this explanation workflow. Uploading that document separately uses the applicable upload workflow.


Privacy Sharing Studio and In-Chat Privacy

Standalone Privacy Sharing Studio helps you obscure sensitive information in supported text, images, and PDFs on your device and prepare a separate sharing copy. The original remains unchanged. Studio does not itself send the content to AI. Review the final output before saving or sharing it. The app applies protection and cleanup mechanisms to its temporary exports; copies you save or share elsewhere are outside that temporary-file lifecycle.

In-chat Privacy creates a new AI request from the cleaned text and attachments you have reviewed. Old chat or workspace file references and unreviewed additional context are excluded from that cleaned request. Accepting the final review item can allow sending to continue without a second “Send” confirmation. The request still uses your existing account authentication and the server-based AI service.

Privacy tools do not create an anonymous account, guarantee zero server or provider retention, or automatically protect every Case Workspace, Library, or other upload. Automated detection can miss identifiers, including information in filenames, document contents, or metadata. Review before sharing.


Medical Images, PDFs, and Slides

Generation workflows can process prompts, reference images and documents, page content, visual instructions, and speaker notes. Document Studio prepares an editable chat draft; when you send it, generation proceeds through the normal AI and backend workflow. Reference materials and generated outputs are therefore part of that processing. These are not exclusively on-device PDF or PowerPoint generation tools, and output capabilities depend on the selected model and tools.


Service Providers

In addition to AI providers, the mobile application integrates Firebase Analytics and Crashlytics for analytics and diagnostics, AppsFlyer for attribution and events, RevenueCat and Apple StoreKit for subscriptions, OneSignal for notifications and related identifiers, and Google Sign-In for optional authentication. Actual processing depends on configuration, feature use, and applicable permissions. Integration of a service does not mean that every user shares every category of information with it.

Web sign-in uses Apple or Google authentication. Web subscriptions use the web payment channel, identified on our pricing pages as Stripe; app-store purchases follow the relevant store’s payment process. Hosting, support, search, and other technical services may also process information needed to provide their functions. Providers’ roles and obligations depend on the service and applicable terms.

We require third parties with whom we share personal information to provide the same or an equivalent level of protection as described in this Privacy Policy.

Official API use alone does not establish a universal “no training” or “zero retention” guarantee. Provider processing, retention, and any model-improvement use depend on the applicable API terms, account settings, and agreements. Do not assume that settings for a provider’s consumer chatbot apply to MedSafeAI requests.


Website Cookies and Similar Technologies

The website uses technologies for sign-in, language preferences, analytics, and marketing attribution. These can process a pseudonymous identifier, campaign parameters, advertising click identifiers, referring domain, and page interactions. The website integrates Google Tag Manager for analytics and measurement tags.

You can manage cookies through your browser and relevant device controls. Blocking cookies can affect sign-in, preferences, or other functionality. Mobile notification and authentication choices are separate from website cookie controls.


Retention and Deletion

Information can be retained to support saved chats, workspaces, files, accounts, subscriptions, support, security, and applicable legal requirements. The relevant period depends on the data category, feature, and operational or legal purpose; there is no single session-only deletion rule for the service.

Where available, you can delete chats or files and request account deletion. In the mobile app, the action labelled “Data Reset” submits an account-deletion request and resets local settings after a successful response; it is not merely a local cache-clearing action. These actions have different scopes: deleting a chat is not the same as deleting an account. Manage subscription cancellation through the original purchase channel; do not rely on app or account deletion to cancel a subscription.

Logs, backups, transaction records, and provider-held copies may follow separate retention and deletion processes. Contact us to request deletion or clarify its scope; we do not promise that every copy is immediately erased.


Security and International Processing

The service uses HTTPS connections and authentication for account-based requests. Privacy Sharing Studio applies protection to its temporary export files. No transmission, storage system, or automated redaction method can be guaranteed completely secure. Protect your device and credentials, and avoid submitting information beyond what your task requires.

Cloud services and third-party providers may process information in countries other than your own. Processing locations and transfer arrangements depend on the service and provider. This policy does not promise storage in a particular country or regulatory certification. Contact us before use if your organisation requires specific hosting, transfer, security, or contractual arrangements.


Your Responsibilities and Privacy Choices

Use MedSafeAI only within your professional authority and applicable confidentiality, privacy, and institutional requirements. Do not submit patient-identifiable information. Remove identifiers and check filenames, metadata, and document contents before submission. You must also have the necessary rights and lawful authority to submit any other clinical or sensitive content, and the intended processing must be permitted. Privacy tools do not replace these responsibilities.

Depending on applicable law, you may have rights to request access, correction, deletion, restriction, objection, or a copy of personal information, and to withdraw consent where processing relies on it. Contact support@medsafeai.com to make a request. We may need to verify your identity and the scope of the request. Do not include patient records or unnecessary sensitive information in your initial message. You may also have the right to contact the relevant data-protection authority.


Changes to This Privacy Policy

We may update this policy to reflect changes to the service or applicable requirements. The revised version and update date will be made available through our legal pages or application. Where required, we will provide additional notice or obtain consent.


Contact Us

SAFE AI NETWORK PTE. LTD.
Email: support@medsafeai.com
Website: https://medsafeai.com
Privacy Policy: https://medsafeai.com/en/privacy