Semantic Privacy Policy

1. Who we are and what this policy covers

Semantic is a software platform that helps teams plan, write, review, publish and measure web content with AI agents. Our customers are businesses: in-house marketing teams running one brand or several, agencies, and other organizations that manage content.

This policy covers:

• The website at semantic.io and its subdomains, except the app (the "Site").
• The cloud service at app.semantic.io, which we host and operate (the "Service").
• Our sales, support, email and other business communications.

This policy does not cover:

• The self-hosted edition of Semantic. When a customer installs Semantic in its own cloud account, that customer runs the software and controls the data in it. We do not receive that data unless the customer turns on optional usage and error reporting (described in section 9) or grants our team access for support.
• Third-party services our customers connect, such as AI model providers, SEO tools, search APIs and content management systems. Their own privacy policies apply. See section 5.

Our two roles. For information about visitors to the Site, prospective customers and the people who hold accounts with us, we decide how the information is used (a "controller" under laws such as the GDPR). For the content and data our customers put into or connect to the Service ("Customer Data"), we process it on our customer's behalf and under its instructions (a "processor" or "service provider"). If you use Semantic through your employer or a client, that organization controls Customer Data, and you should direct questions about it to them first. Our customer agreement and data processing addendum govern how we handle Customer Data and take precedence over this policy where they differ.

2. Information we collect

Information you give us

• Account information: your name, email address, organization, role and the brands you belong to. If you sign in with a password, we store only a one-way hash of it. If you set up two-factor authentication, we store its secret encrypted.
• Sign in with Google: if you choose it, Google shares your name, email address, profile picture and Google account ID with us. We use them only to create and secure your Semantic account.
• Invitations: when an Admin invites a teammate, we receive that person's email address and the role offered.
• Customer Data: what your team adds to the Service, such as brand guidelines and voice, product and proof details, uploaded documents, briefs, drafts, comments, approvals, evidence and sources, content calendars, keyword lists, and settings.
• Credentials for connected services: API keys and logins your team enters for AI model providers, search, SEO and publishing tools. These are stored encrypted (section 10).
• Billing information: billing contact details and payment records. Card payments are handled by Stripe; we do not store full card numbers.
• Communications: what you send us by email, through support, or through forms on the Site, such as a demo request or newsletter signup.

Information from connected services

When your team connects a service to a brand, the Service reads data from it to provide the features you asked for. Examples: Google Search Console and Google Analytics (section 3), your website's pages and sitemaps (through our crawler), your content management system (WordPress, Webflow, Shopify and others), SEO tools such as Semrush and Ahrefs, and chat tools such as Slack or Microsoft Teams for notifications.

External reviewers

A customer can send a review link to someone outside its Semantic account, such as a subject-matter expert or a client. When that person opens the link, we record the name and email address they enter, their comments and approval decisions, and the time and IP address of each action.

Information collected automatically

• In the Service: IP address, browser and device type (user agent), session records, sign-in events, and a log of actions taken in each brand (who did what, when, and from which session). This audit log is a core security and accountability feature that customers rely on.
• On the Site: cookies, pixels and similar technologies (section 6).

What the Service is not built for. Semantic works with marketing content, keyword data, aggregate website analytics and account information. It is not designed to process your own customers' personal information, and Google Analytics data in Semantic is aggregate (for example, sessions per landing page). Please do not upload sensitive personal information, such as health or financial records, to the Service.

3. Google user data

This section explains exactly how Semantic accesses, uses, stores, shares and deletes data it receives from Google APIs.

Limited Use. Semantic's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.

What we access, and only when you choose to connect

Semantic asks only for view-only access. It cannot change your Search Console or Analytics settings, add or remove users, or delete data in your Google accounts. Google Analytics data in Semantic is aggregate; we do not request user-level or visitor-level reports.

How we use it. Only to provide and improve the user-facing features of the brand you connected, such as:

• showing search and traffic performance on dashboards, the content inventory and page details;
• keeping a daily history of the brand's data beyond Google's own 16-month window, so before-and-after comparisons work for older changes;
• labeling pages that are declining, gaining, close to page one or not indexed, and suggesting next steps such as a content refresh;
• giving the brand's AI agents the numbers they need to write a brief, a refresh recommendation or a report summary that you asked for; and
• sending the scheduled reports your team sets up.

Who we share it with. We do not sell Google user data, and we do not use it or share it for advertising, retargeting, interest-based advertising, credit decisions or lending, or with data brokers or information resellers. We share it only:

• with the people your organization has given access to that brand in Semantic, according to the roles it assigns, and with the recipients of reports your team schedules;
• with the AI model provider your organization has connected to the brand (for example Anthropic, OpenAI, Google or xAI), through your organization's own provider account and key, and only the data needed for the feature you asked for (section 5);
• with our infrastructure providers, such as Amazon Web Services, which store and process it on our behalf under confidentiality and security obligations (section 7); and
• when required by law, to protect the security of the Service, or as part of a merger or acquisition, and in that last case only with your prior consent.

No training of AI models. We do not use Google user data to develop, improve or train generalized or non-personalized AI or machine learning models, and we do not permit anyone to whom we transfer it to do so. When an AI model provider processes Google user data through your organization's account, it does so to produce output for your brand.

When people at Semantic can see it. Our staff do not read your Google user data unless: (a) your organization gives us permission, for example by granting our team operator access to a brand or asking us for support with a specific issue; (b) it is necessary for security purposes, such as investigating abuse or a security incident; (c) it is required to comply with law; or (d) it has been aggregated and anonymized for our internal operations.

How we protect it. Google user data is encrypted in transit (TLS) and at rest. Google access tokens are stored encrypted and are used only by the back-end services that call Google. Each brand's data is isolated from every other brand's data by database-level controls (section 10).

How long we keep it, and how to delete it

• Google access tokens are kept until you disconnect Google in the brand's settings or revoke Semantic's access in your Google Account. When you disconnect, we delete the tokens and stop reading data.
• Search Console and Analytics history stays with the brand while it is active, so its reports keep working. A brand Admin can delete it at any time in settings. It is also deleted when the brand or your organization's account is deleted (section 8). Deleted data is removed from our backups within 35 days.
• Sign in with Google profile data is kept for as long as your Semantic account exists, and is deleted when the account is deleted.
• You can ask us to delete Google user data at any time by emailing semanticadmin+privacy@gmail.com. If your organization manages your account, we may need to confirm the request with its Admin.

Optional indexing assistant. Google offers no API for the Request indexing button in Search Console. A brand Admin can turn on an optional, off-by-default feature in which a person signs in to a Google account once inside an isolated cloud browser that belongs to that brand, and Semantic's agent then requests indexing for the brand's own new and updated pages. Semantic never sees or stores the Google password. The browser visits only Search Console pages for the brand's own properties, its session is stored encrypted and separate from every other brand, and everything in this section applies to what it sees. An Admin can turn the feature off and sign the browser out at any time, which deletes the stored session.

4. How we use information

Customer Data is used only to provide the Service to that customer. We do not use Customer Data, including your content, brand guidelines and connected analytics, to train AI models, and we do not sell it or use it to advertise to anyone. We may use aggregated usage statistics that do not identify any customer, brand or person (for example, the average time a draft waits for review) to run and improve the Service. Customer Data that came from Google APIs is also subject to section 3.

5. AI model providers and other services your organization connects

Semantic runs on your organization's own accounts with the services it uses ("bring your own key"). An Admin adds each key to a brand, and the Service uses that key on that brand's behalf. This means:

• AI model providers. When an agent drafts, edits, reviews, researches or summarizes, Semantic sends the relevant brand content and instructions to the model provider your organization connected, such as Anthropic, OpenAI, Google (Gemini) or xAI. Image generation works the same way. The provider processes that data under your organization's agreement with it, not ours. Check that provider's terms, including whether it may retain or train on what it receives.
• Web search. When an agent researches a topic, it sends search queries (not your drafts) to the search service your organization connected: Brave Search, DataForSEO, or the AI model provider's built-in search.
• SEO tools such as Semrush and Ahrefs receive the domains, URLs and keywords Semantic asks them about.
• Publishing tools. When your team approves a piece, Semantic sends it to the content management system your organization connected (for example WordPress, Webflow or Shopify). Where a CMS has no API, an optional browser agent signs in to it in an isolated, single-use browser.
• Chat and notifications. If your team connects Slack or Microsoft Teams, Semantic sends notifications and report summaries there.
• AI-answer visibility. If your team sets up a prompt panel, Semantic sends those prompts to the AI providers your organization connected and records whether the answers mention or cite the brand.

We choose what each request contains, and we send only what the task needs. Credentials for these services are stored encrypted and are never shown again after they are saved, including to our staff. Because your organization chooses these services and contracts with them directly, they are not our subprocessors, and we are not responsible for their practices.

Prompts and fetched web content are treated as untrusted. Our agents are built to ignore instructions hidden in web pages, documents or search results.

6. Cookies, pixels and your choices

In the app (app.semantic.io) we use only cookies that the Service needs to work: to keep you signed in, protect against forged requests, and remember settings such as your active brand. The app contains no advertising pixels or third-party analytics tags.

On the Site (semantic.io) we use Google Tag Manager to load the tools below. Tag Manager itself does not collect personal information; it decides which tags may run based on your consent choice.

These tools may collect your IP address, browser and device information, pages visited, referring site, the time of your visit, and identifiers stored in cookies. Meta, LinkedIn and X may link this to an account you hold with them, under their own privacy policies.

How consent works where you are

• If you are in the United States, analytics and advertising tags run when you arrive, and you can opt out at any time with the Your Privacy Choices link in the Site footer. We also treat a Global Privacy Control (GPC) signal from your browser as an opt-out. Opting out turns off the advertising tags for that browser.
• If you are anywhere else, including the European Economic Area, the United Kingdom and Switzerland, only necessary tools run until you choose otherwise in our cookie banner. Analytics and advertising tags run only after you accept them. We use Google Consent Mode, so Google tags respect your choice.
• You can change or withdraw your choice at any time through Cookie settings in the Site footer, and you can also block or delete cookies in your browser settings.

The Site does not respond to the older "Do Not Track" browser setting, because there is no common standard for it; we honor GPC instead.

7. How we share information

We do not sell Customer Data or account information. We share information only as follows.

• Within your organization. People your organization adds to a brand see that brand's data according to the role an Admin gives them. Your organization's Admins can see the audit log, which records who did what in the brand.
• With external reviewers your team invites. A review link shows the reviewer the content your team chose to share, and your team sees the reviewer's comments and decisions.
• With our team, when your organization grants operator access. Some customers ask Semantic's team to run content work for them. An Admin grants this per brand, chooses the role, and can set an expiry and revoke it at any time. Our operators appear in the brand's user list with an Operator badge, and every action they take is logged under their own name.
• With services your organization connects (section 5), when the Service uses them on your behalf.
• With our service providers (subprocessors), who process data on our behalf under contracts that require confidentiality and security:

• With advertising and analytics partners on the Site (section 6), subject to your consent choices.
• For legal reasons, when we believe in good faith that disclosure is required by law or legal process, or is needed to protect the rights, property or safety of Semantic, our customers or others. Where the law allows, we will tell the affected customer before disclosing its Customer Data.
• In a business transfer, such as a merger, acquisition or sale of assets, subject to this policy. For Google user data, we will ask for your consent first (section 3).
• With your consent or at your direction.

We will keep an up-to-date subprocessor list and give customers 30 days' notice before adding one.

8. How long we keep information

Deleted data is removed from our backups within 35 days, as the backups expire.

9. The self-hosted edition

In the self-hosted edition, the customer runs Semantic in its own cloud account and is responsible for the data in it. Usage and error reporting to Semantic is off by default. If the customer turns it on, we receive technical information such as the software version, feature usage counts and error reports, which are designed to exclude Customer Data. If a customer invites our team to its installation for support, we access only what that customer grants.

10. How we protect information

• Data is encrypted in transit (TLS) and at rest.
• API keys, OAuth tokens and other secrets are encrypted with per-brand keys that are themselves protected by AWS Key Management Service. They are decrypted only by the back-end service making the call and never appear in the browser, logs or AI prompts.
• Each brand's data is isolated by row-level security in the database and separate storage prefixes. Automated tests try to read across brands through every part of the API and must be refused.
• Sign-in supports Google, one-time email links, and passwords with mandatory two-factor authentication. Admins can require single sign-on.
• The audit log is append-only and tamper-evident.
• Staff access to production systems is limited to people who need it, and is logged.

No system is perfectly secure. If we learn of a security breach affecting your information, we will notify affected customers and individuals as the law requires.

11. International transfers

We are based in the United States, and the Service is hosted in the United States. If you are outside the US, your information will be transferred to and processed in the US. Where the law requires it, we use safeguards such as the European Commission's Standard Contractual Clauses and the UK International Data Transfer Addendum.

12. Your rights and choices

Everyone. You can update your account details in the app, unsubscribe from marketing email with the link in each message, and change your cookie choices on the Site at any time.

Depending on where you live, you may have the right to access, correct, delete or receive a copy of your personal information, to object to or restrict certain uses, to withdraw consent, and to complain to your local data protection authority. To use these rights, email semanticadmin+privacy@gmail.com. We will confirm your identity before acting and respond within the time the law requires. If your request concerns Customer Data that an organization controls, we will refer you to that organization or help it respond.

US state privacy rights. Residents of California and other states with privacy laws may ask what personal information we collect, use and disclose; ask us to delete or correct it; and opt out of the "sale" or "sharing" of personal information or its use for targeted advertising. We do not sell personal information for money. The advertising tags on our Site (section 6) may count as "selling" or "sharing" under these laws, and you can opt out with the Your Privacy Choices link or a Global Privacy Control signal. We do not use sensitive personal information to infer characteristics about you. In the past 12 months we collected the categories described in section 2: identifiers, internet and network activity, commercial information (billing), and professional information. You may use an authorized agent, and you may appeal our response to a request by replying to it. We will not discriminate against you for using these rights.

13. Children

Semantic is a business service for adults. It is not directed to anyone under 18, and we do not knowingly collect information from children. If you believe a child has given us information, contact us and we will delete it.

14. Changes to this policy

We will post any changes here and update the date at the top. If a change is material, we will tell account holders by email or in the app before it takes effect. We will not use Google user data in a new way without first telling you and, where required, asking for your consent.

15. Contact us