Privacy Policy for Exmergo Viz

Last modified: August 16, 2026

1. Scope

This annex describes how Exmergo, Inc. handles data in Exmergo Viz, our hosted analytics and dashboard service at viz.exmergo.com.

It forms part of the Exmergo Privacy Policy and should be read with it. The main policy covers who we are, your rights, international transfers, security, retention of account records, and how to contact us. This annex covers what is specific to Viz. Where the two differ on a Viz question, this annex governs.

Exmergo Dex is a different product with a different architecture and is covered by its own annex, the Privacy Policy for Exmergo Dex.

2. Our Role for Exmergo Viz

• We are a processor for the content you put into Viz: the files you upload, the sources you connect, the prompts you write, and the dashboards you build. We process that content on your instructions, for your purposes. If it contains personal data about other people, you are the controller of it.
• We are a controller for your account information, your billing records, and the usage and analytics data we collect about how you use the service.

A Data Processing Agreement covering our role as processor is available on request at legal@exmergo.com.

3. What We Collect in Exmergo Viz

A. Information you provide

• Account information. When you create an account through Google or Microsoft, we receive your name, email address, and profile picture from that provider. We store this alongside your Viz settings, including your pricing tier.
• Guest accounts. If you use Viz without signing up, we create an anonymous account so your work persists within the session. If you later sign up, we may use a temporary migration key to attach that work to your new account. The key is short-lived and exists only for that transfer.
• Uploaded data. The files you upload for analysis, in formats including CSV, XLSX, JSON, and Parquet. We do not inspect, restrict, or catalogue the contents of these files beyond what is needed to query them for you.
• Connected sources. Where your plan supports it, the credentials and connection settings for a data source you choose to connect. See section 7.
• Prompts. The questions and instructions you type into the chat interface.
• Payment information. Collected directly by Stripe, Inc. We never receive or store your full card number.

B. Information generated as you work

• AI-generated content. The queries and visualization code our agents produce in response to your prompts.
• Dashboard and session data. The configuration of your dashboards, including chart titles, sizes, and positions, and the history of your conversation with the AI for each session.

C. Information collected automatically

• Usage and analytics data, collected through Google Analytics: device and browser type, operating system, screen resolution, approximate location, how you arrived, which features you used, and session duration, with a randomly generated client identifier stored in a cookie.
• Server logs, recording IP address, request details, and timestamps for security and debugging.

4. How Your Data Is Stored and Queried

• Conversion and storage. Files you upload are converted to Apache Parquet and stored privately on Google Cloud Platform.
• Segregation. Each file is written to a storage path unique to your user ID and session ID. Your data is not co-mingled with any other customer's, and no query path exists from one account to another's storage.
• Querying. Our data engine reads your files directly from that private location. Query results are computed inside our infrastructure and returned to your browser.
• Encryption. Data is encrypted in transit over HTTPS and at rest.

5. What the AI Models Receive

This is the part of Viz that most people want a precise answer on, so here it is precisely.

We do not send your raw data file to Google or to any other AI provider. The file stays in our storage and is queried by our own engine.

To generate a relevant query or visualization, our agents send Google's Gemini models a bounded subset:

1. Your conversational prompt, for example "show me sales by region."
2. The conversation history for that session.
3. The schema of your data: column names and data types.
4. A random sample of rows, so the model can tell what the columns actually contain.
5. Aggregated results computed from your data, so the model can describe patterns and write insights.

Items 4 and 5 mean that some values from your data do reach the model. We say so plainly rather than claim a boundary we do not hold. What the model never receives is the complete dataset, the underlying file, or your stored credentials.

If your data is sensitive enough that a sample of rows must not leave our infrastructure, an Enterprise deployment is the right fit. See section 12, and speak to us before you upload.

6. AI Model Training

• Free plan. We may use your conversational prompts, data schema, and data samples to train and refine the performance of our AI agents. Choosing the free plan is how that consent is given, and you can withdraw it at any time by upgrading to a paid plan or by deleting your account.
• Pro, Ultra, and Enterprise plans. Your data, prompts, and schema are never used to train our AI models. This is not a setting you need to find and switch off. It applies from the moment your paid plan begins.

Separately from either case, we analyze usage data in aggregated and anonymized form to understand which features are used and to improve the product. That analysis cannot be traced back to an individual user or dataset.

7. Connected Data Sources and Their Credentials

Beyond file upload, which is available on every plan, Viz can connect directly to systems you already use. Availability depends on your plan:

When you connect a source, we store the connection settings and the credentials or access tokens needed to reach it. Those are encrypted at rest, are scoped to your account, and are never sent to an AI provider, never written into a dashboard, and never exposed to anyone you share a dashboard with. Connections are encrypted in transit. You can revoke a connection at any time from within Viz, which deletes the stored credential.

Enterprise deployments can additionally place these connections behind VPN or VPC access controls.

8. Published and Shared Dashboards

Publishing a dashboard creates a read-only public snapshot. It does not make your uploaded file public.

When you publish, our engine re-runs the specific queries behind each visualization on your dashboard and saves only the final results to a separate public location. Anyone with the link can view the resulting charts. They cannot access, query, download, or reconstruct the underlying dataset, and they cannot see any column you did not chart.

Treat a published link as public. Anyone who has it can open it, so publish only aggregates you are comfortable making public, and unpublish from within Viz when you no longer want the link to work.

9. Retention and Deletion

• Uploaded files. Retained while your account is active. Deleting a file inside Viz permanently deletes it from our systems.
• Dashboards and session history. Retained while your account is active, and deletable individually from within Viz.
• Published snapshots. Retained until you unpublish the dashboard or delete it.
• Account, billing, and correspondence records. Retained as described in section 10 of the main policy.

Deleting your account removes your uploaded files, dashboards, and session history. We may retain billing records where accounting or tax law requires it.

10. Where Your Data Is Stored

• Free, Pro, and Ultra plans. All data is processed and stored in the United States. If you are outside the United States, using these plans means your information is transferred there.
• Enterprise plans. You may select a geography, for example within the European Union, for storage and processing. Your enterprise agreement sets the specifics.

Transfers out of the EEA, the UK, or Switzerland rely on the Standard Contractual Clauses, as described in section 12 of the main policy.

11. Sub-processors for Exmergo Viz

• Google Cloud Platform and Firebase. Hosting, file storage, the application database, and analytics.
• Google (Gemini). The AI models, receiving only what section 5 describes.
• Google and Microsoft. Authentication, when you sign in through them.
• Stripe, Inc. Payment processing on paid plans.

We do not sell your personal data and we do not share it for advertising.

12. Enterprise Deployment Options

Enterprise customers can change where Viz runs, which changes the answer to several of the questions above:

• Standard Cloud. Shared infrastructure with logical isolation, the same architecture described throughout this annex.
• Private Managed Cloud. Dedicated infrastructure with physical isolation and a geography of your choosing.
• On-Premise. Viz runs inside your own environment. Your data does not reach Exmergo infrastructure at all, and our role narrows to the account and licensing information you provide us directly.

Your enterprise agreement, and any Data Processing Agreement executed with it, governs where these differ from this annex.

13. Prohibited Data

As set out in our Terms of Service, you may not upload data subject to specific regulatory regimes, including Protected Health Information under HIPAA or cardholder data under PCI-DSS, unless you have a separate written agreement with us covering it. Viz is not configured for those regimes by default, and uploading such data without that agreement puts it outside the protections either of us has committed to.

14. Contact

Questions about this annex, or requests to exercise your rights, go to support@exmergo.com. Data Processing Agreements and contractual questions go to legal@exmergo.com. The full contact details are in section 16 of the main policy.