Privacy Policy

At a glance

• Personal data we collect: your name and email; the documents, teams, and vaults you create; records of the AI clients you connect (tokens stored only as hashes); an activity log of actions in your vault; billing identifiers from Stripe (never card details); IP address and request metadata; and browser usage data collected by the analytics services described below.
• Why we use it: to provide and secure the Service, enforce team and vault permissions, prevent abuse, show usage insights inside your own vault, understand product usage, measure how well our own ads work, process payments, and send transactional and account email. We do not sell personal information and do not show third-party advertising.
• Who receives it: only the service providers that run Verde — Neon (database and sign-in), Vercel (hosting, bot checks on the sign-up and sign-in pages, and cookieless page-view counts), Stripe (billing), Resend (email), Google (sign-in with Google, bot checks on the sign-up and sign-in pages, and analytics on public pages only), Meta (measuring our own ads: our marketing and sign-in pages, and the sign-ups that came through them), OpenAI (measuring our own ads in ChatGPT: the sign-ups that came from them), Microsoft Clarity (usage analytics and session replay), and Sentry (error monitoring) — plus AI clients you authorize and people with access to your shared or published content. We do not use your content to train AI models. Connected AI clients receive data through the integration as described below.
• How long we keep it: your content for as long as its team or vault exists; activity-log entries for your plan’s window (30 days on Free, 90 on Team, 365 on Business, unlimited on Enterprise); rate-limit records for one window; a deleted account’s data for a 30-day restore period, then permanent deletion or anonymization.
• Your controls: export your data as JSON, edit or delete your content, delete teams and vaults you own, turn off account email, disconnect AI clients at any time, and delete your account from Settings. Turning on Global Privacy Control in your browser switches off our ad measurement. GDPR, UK GDPR, and CCPA rights are honored on request.
• When you connect an AI client (for example ChatGPT, Claude, or Cursor), it reads and writes your documents over MCP with exactly your permissions. What that client then does with the content is governed by its own provider’s terms, not by us.

Information we collect

• Account information. Your name and email address, managed through our authentication provider (Neon Auth / Better Auth). If you sign in with Google, we receive basic profile information (name, email) from Google.
• Content you create. The teams, vaults, buckets, and documents (“memories”) you and your teammates create, including their text, metadata, tags, comments, revision history, and typed document properties such as a task state, priority, date, or assigned team member. Person properties may be supplied by name, email, or member identifier and are stored as a reference to that team member.
• Connections and tokens. When you connect an AI client, we store records of the OAuth clients you authorize and hashed representations of access and refresh tokens (the plaintext is shown only once at creation and is not stored).
• Activity data. An activity log of actions taken in your vault (for example, documents created or searches run) to power usage insights and audit history. Entries record the actor, the action, whether it came from the dashboard or an AI client, and (for searches) the query text.
• Billing information. If your team subscribes to a paid plan, we store your plan, subscription status, billing interval, seat count, and the identifiers Stripe assigns to your customer and subscription records. We send Stripe your team name and the billing contact’s email address. Card details are entered directly into Stripe and never reach our servers; we do not store card numbers, expiry dates, billing addresses, or tax identifiers.
• Technical data. IP address and basic request metadata, used for security, abuse prevention, and rate limiting. Rate limiting stores a short-lived record derived from your IP address in our database; it expires as each rate-limit window rolls over.
• Browser usage data. The analytics services described below receive page addresses, browser and device information, and interaction data. Microsoft Clarity can also receive page content for session replay, including content in the signed-in dashboard, depending on the configured masking settings.

Vault activity and rate-limit records. The activity log exists to show your own team its usage history, is visible only to members of that vault (entries about your private documents, and searches limited to them, are visible only to you, and the latter are logged without their query text), and expires on your plan’s retention window; the rate-limit record expires within minutes. Deleting a team or vault removes its activity log, and deleting your account strips your identity from any entries that remain after the deletion grace period. Browser analytics and error-monitoring data are separate from these vault records.

Restricted data we do not collect. Verde does not request protected health information, government identifiers such as social security numbers, or credentials and authentication secrets for other services. Payment card details go directly to Stripe and never reach us. Documents can contain information you choose to submit, so do not include these sensitive categories. Our own access tokens are stored only as hashes, and the guidance we give every connected AI client instructs it never to save passwords, API keys, or other secrets as documents. Please do not put such information in your documents.

Cookies and local storage

Our authentication provider sets cookies that are strictly necessary to sign you in and keep you signed in. The analytics services below also use cookies and similar technologies. Your browser stores a small amount of local preference data (such as your light/dark theme choice and whether you have dismissed onboarding tips); that data stays in your browser.

Our public pages (the marketing site, sign-in pages, and public wikis) use Google Analytics to measure visits, which sets analytics cookies and shares page addresses and technical data (such as IP address and browser type) with Google. Google Analytics never runs inside the signed-in app: the pages, teams, vaults, and documents you work with are not reported to Google.

Every page, including the signed-in app, uses Vercel Web Analytics to count page views. It sets no cookies and stores nothing in your browser. It receives the page address, the site that referred you, and browser, device, and approximate location information. Before anything is sent, the address is reduced to the kind of page (such as a vault’s settings page), with team, vault, and document names and identifiers and invitation and share links removed, and its query string is removed except for campaign tags. Only our public wiki and blog pages are reported by their full address.

The sign-up and sign-in pages check that a person, not an automated program, is entering an email address or creating an account. They use Google reCAPTCHA, which runs on those pages only, sets cookies, and collects hardware and software information, such as device and application data, and IP address, and sends it to Google to provide and improve reCAPTCHA and for general security. Google does not use it for personalized advertising and handles it under its Privacy Policy. Vercel BotID, part of our hosting, runs a similar check on the same forms. We keep the reCAPTCHA score with the new account.

The same marketing and sign-in pages, but not the public wikis, use the Meta Pixel to measure how well our ads on Facebook and Instagram work, such as whether a visit that began with one of our ads led to a sign-up. It sets cookies and shares the page address, the referring page, browser and device information, IP address, and Meta cookie identifiers with Meta, which may link them to a Meta account you hold and handles them under its Privacy Policy. The Meta Pixel never runs inside the signed-in app, never reports a page whose address carries account details (such as an invitation or password-reset link), and does not load at all when your browser sends a Global Privacy Control signal.

When you create an account in a browser where the Meta Pixel has already run, or that arrived from one of our Meta ads, our server also tells Meta about the sign-up, once, so Meta can match it to the ad. It sends a hashed (SHA-256) copy of your email address, your IP address, your browser’s user agent, and the Meta cookie identifiers. Nothing is sent for a browser that sends Global Privacy Control, or for a sign-up that never met the Meta Pixel or one of our ads, such as one made straight from a team invitation.

When a visit begins with one of our ads in ChatGPT, OpenAI adds a click reference to the page address. Our marketing and sign-in pages keep that reference in a first-party cookie, __oppref, for 30 days. Your browser does not send it to OpenAI, and we do not load OpenAI’s measurement pixel. If you create an account in that browser while the cookie is present, our server tells OpenAI about the sign-up, once, so OpenAI can match it to the ad. It sends the click reference, a hashed (SHA-256) copy of your email address, your IP address, your browser’s user agent, and a random identifier for your account. OpenAI handles this data under its Privacy Policy. Nothing is stored or sent for a browser that sends Global Privacy Control, or for a sign-up that did not come from one of our ChatGPT ads. This is separate from connecting ChatGPT or Codex to Verde as an AI client, described below.

Microsoft Clarity is enabled on public pages and the signed-in dashboard to help us understand product usage through heatmaps and session replays. It collects interaction and technical data, such as clicks, scrolling, page addresses, browser information, and session identifiers, and may use cookies. Page text, including document text and vault names, may be included where it is not masked by our Clarity configuration. Clarity is a browser integration; it does not run on MCP API requests. Microsoft handles this data under its Privacy Statement.

Connected AI clients and Automatic Memory

When you connect ChatGPT, Codex, Claude, or another AI client, you authorize it to access Verde through MCP within your account, team, vault, and document permissions. Verde receives the arguments that client sends in tool calls, such as search queries, document content, comments, property values, and member names or emails used to resolve person fields. We do not receive your entire conversation through this integration; we receive the information included in those calls.

Tool responses can return accessible document text or excerpts, titles, tags, properties, comments, author and assignee names, document and vault identifiers, links, timestamps, versions, and publication or visibility status. These fields let the client find, cite, organize, and update the right document. Person properties normally return the member’s name; an unresolved stored member identifier may be returned instead. Data returned to a connected client is shared with its provider, including OpenAI when you use ChatGPT or Codex, and is subject to that provider’s terms and your settings with it. Verde does not use your content to train AI models; search runs inside our database.

Automatic Memory is a per-connection setting, enabled by default on new connections with a visible checkbox you can turn off when connecting. When enabled, an AI client may save durable information from your conversation without asking before each save. Those saves become your private drafts, and the client is instructed to tell you what it saved. Publishing an automatically saved draft requires explicit approval after you have reviewed its content. You can turn Automatic Memory off or revoke a connection from Verde’s MCP Connections settings. Revoking access stops future calls but does not delete previously saved documents or copies already received by the AI provider; those follow the respective retention and deletion rules.

Company documents are available to authorized vault members; private documents are limited to their owner. Publishing a public document makes its public page accessible without sign-in. A share link makes its document accessible to people who have that link until it expires or is revoked. Document properties and comments are not displayed on public wiki or shared pages, although a property update changes the modification date shown there.

How we use information

• to provide, maintain, and secure the Service;
• to enforce permissions and team and vault boundaries;
• to prevent abuse (including rate limiting) and diagnose errors;
• to show usage insights within your own vault;
• to understand browser interactions and improve the product;
• to measure how well our own ads work;
• to process payments, manage subscriptions and seat counts, and enforce plan limits;
• to send transactional email: team invitations, sign-in codes, and password-reset links;
• to send account email about your own use of the Service: a welcome when you sign up, a few tips while you get started, a reminder when an upgrade is left unfinished, and a weekly digest of your teams’ activity. You can turn these off from the link in each one or in Settings. If someone invites you to a team, we may send you one reminder if you haven’t accepted after a few days. We count clicks on the links in these emails, not opens, to learn which ones are useful. We do not send third-party advertising.

We do not sell your personal information. Our ad measurement with Meta and OpenAI (see “Cookies and local storage”) may count as “sharing” for cross-context behavioral advertising under California law. You can opt out by turning on Global Privacy Control in your browser, which we honor, or by contacting us.

Service providers (sub-processors)

We share data with vendors that help us run the Service, only as needed to provide it:

• Neon: managed Postgres database and managed authentication (identity and sessions). Some authentication email, such as email verification, may be delivered by Neon’s own email provider.
• Vercel: application hosting and serverless compute, Vercel BotID bot checks on the sign-up and sign-in pages, and Vercel Web Analytics page-view counts on every page (receives the reduced page addresses and browser information described in “Cookies and local storage” above).
• Cloudflare: private object storage (R2) for files your team uploads to a vault on a paid plan. Files are stored encrypted at rest, are never given a public address, and are reachable only through short-lived links Verde issues after checking the requester’s access.
• Stripe: payment processing and subscription billing for paid plans (receives your team name, billing contact email, and card details entered directly into Stripe’s payment form).
• Resend: delivery of email, such as team invitations, sign-in codes, and the account email described above (receives the recipient address and the contents of that email).
• Google: sign-in, if you choose to sign in with Google; reCAPTCHA bot checks on the sign-up and sign-in pages; and Google Analytics on our public pages only (see “Cookies and local storage” above).
• Meta: measurement of our own ads on Facebook and Instagram, through the Meta Pixel on our marketing and sign-in pages only, and a sign-up event from our server for sign-ups that came through them (receives a hashed email address, IP address, user agent, and Meta cookie identifiers; see “Cookies and local storage” above).
• OpenAI: measurement of our own ads in ChatGPT, through a sign-up event from our server for sign-ups that came from them (receives the ad’s click reference, a hashed email address, IP address, user agent, and a random account identifier; see “Cookies and local storage” above). This is separate from your use of ChatGPT or Codex as a connected AI client.
• Microsoft Clarity: browser usage analytics, heatmaps, and session replay on public pages and the signed-in dashboard (receives the usage data and potentially unmasked page content described above).
• Sentry: error monitoring, if enabled for the deployment (receives error details, stack traces, and limited request context such as the URL path).

We do not use your content to train AI models. Search runs inside our database. When you authorize an AI client over MCP, we return data to it as described in “Connected AI clients and Automatic Memory” above; its provider handles that data under its own terms and your settings with it.

Our database and application are hosted in the United States (AWS US East). If you access Verde from outside the United States, your information will be transferred to and processed there.

Data retention

We retain Your Content for as long as your team or vault exists. Documents are archived rather than erased, so history stays available; deleting a team or vault is immediate and permanent and removes its content, including archived documents (subject to limited backup retention).

Activity log retention depends on your plan. We automatically delete activity-log entries older than your plan’s window (30 days on Free, 90 days on Team, 365 days on Business, and unlimited on Enterprise), with a short grace buffer before deletion runs. If a paid subscription lapses, the team’s activity log is pruned on the Free window, which permanently deletes older entries. This applies only to the activity log; your documents are not deleted by plan retention.

Your rights and choices

• Access / export. You can export a copy of your account and content data as a JSON file from your account settings. The export covers your profile, team memberships, the documents you authored, your access-token records (never the token secrets), and your activity history. It does not include records held by our authentication provider, such as sign-in history. Contact us to request those.
• Account deletion. You can delete your account from your account settings. Deletion is a soft delete with a 30-day grace period: your account is deactivated immediately (you’re signed out and access is blocked), but your data is retained for 30 days so you can restore it by signing back in. After 30 days, your personal data is permanently deleted or anonymized: your private documents and connections are removed, and content you contributed to shared teams is retained but stripped of your identity. If you are the sole owner of a team with other members, you’ll be asked to transfer ownership first so the team isn’t left stranded. Deleting your account also cancels any subscription for teams that are removed with it. One limitation to be aware of: the identity record held by our authentication provider (Neon Auth), such as your email address and sign-in history, is not always removable by our automated purge. Where it cannot be removed automatically, your account remains permanently blocked from signing in, and you can contact us to request erasure of the remaining identity record.
• Team and vault deletion. You can delete teams and vaults you own from their settings at any time.
• Correction. You can edit your content and profile at any time.

Depending on where you live, you may have additional rights under laws such as the GDPR or CCPA. Contact us to exercise them; we do not discriminate against you for doing so.

Business customers and data processing agreements

If you use Verde on behalf of a business and need a Data Processing Agreement, we will provide one on request. Contact privacy@getverde.ai. Where required for transfers of personal data out of the European Economic Area or the United Kingdom, that agreement incorporates Standard Contractual Clauses.

Users in the European Economic Area and United Kingdom

We do not currently market or direct Verde to the EEA or UK, and we have no establishment there. We recognize that people in those regions may nonetheless find and use the Service, so we aim to handle their information consistently with the GDPR and UK GDPR.

Legal bases. Where the GDPR applies, we rely on: performance of a contract to create your account, host your content, and provide the Service; legitimate interests to secure the Service, prevent abuse, diagnose errors, and understand usage within your own workspace; legal obligation to meet tax, accounting, and compliance duties; and consent where we ask for it, which you may withdraw at any time.

International transfers. Our infrastructure is in the United States, so using the Service means your information is transferred there. Where required, we rely on appropriate safeguards for those transfers, including the European Commission’s Standard Contractual Clauses, and our sub-processors are engaged under comparable terms.

Your rights. Subject to the conditions in those laws, you may request access to, correction of, deletion of, or a portable copy of your personal data; object to or ask us to restrict certain processing; and lodge a complaint with your local supervisory authority. Many of these are available directly in your account settings; for the rest, contact us at the address below and we will respond within the time the law allows.

We have not appointed an EU or UK representative under Article 27, as we do not target those markets. We will revisit this if that changes.

Security

Data is transmitted over TLS. Access and refresh tokens are stored only as hashes. Access to your content is enforced server-side by vault, role, permission level, and ownership rules, and is re-checked on every request, including requests made by AI clients, so revoking someone’s access takes effect immediately. No system is perfectly secure, but we work to protect your information.

Children

Verde is designed for teams and business use. You must be at least 13 years old to use it, and users under the age of majority where they live need the consent of a parent, legal guardian, or authorizing organization, as described in our Terms of Service. The Service is not directed to children under 13, we do not knowingly collect personal information from them, and we will delete any such information we discover. If you believe a child under 13 has provided us information, contact us at the address below.

Changes to this policy

We may update this policy; material changes will be reflected by updating the “Last updated” date above.

Contact

Privacy questions or requests: privacy@getverde.ai.