Legal · Privacy

Privacy Policy

This page says what Actionfleet collects, where it goes, and how to get it removed. It describes what the service does today. Actionfleet is pre-alpha, and this policy will grow with it.

Last updated November 10, 2026. Questions go to support@actionfleet.ai.

Contents

  1. 01Who we are
  2. 02What this covers
  3. 03Accounts
  4. 04Google user data
  5. 05Orgs, Fleets and Ships
  6. 06Agent connections and machines
  7. 07Secrets and API keys
  8. 08GitHub and Tailscale
  9. 09The email list
  10. 10Advisory workspace and payments
  11. 11Page views and events
  12. 12Hosting and request logs
  13. 13Cookies and browser storage
  14. 14What we do not do
  15. 15How we use it
  16. 16Who else sees it
  17. 17How long we keep it
  18. 18Your choices
  19. 19Children
  20. 20Changes to this policy
  21. 21Contact

Who we are

actionfleet.ai and the Actionfleet app are run by Actionfleet, a sole proprietorship. When this policy says “we” or “Actionfleet,” it means that company.

The way to reach us about anything on this page is support@actionfleet.ai.

What this covers

This policy covers actionfleet.ai and its subdomains: the public pages, the email list, your account, the Actionfleet app, its MCP server that agents connect to, and the Actionfleet command-line tool on machines you add. If we start collecting something not described here, we update this page first.

Accounts

You can create an account with an email address and a password, or with Google sign-in. Accounts are held in Google Cloud Identity Platform (Firebase Authentication), which we operate inside our own Google Cloud project.

  • Email and password. Your browser sends the password straight to Identity Platform, which stores it as a salted hash. Our own servers never receive the password.
  • Google sign-in. Google asks you to approve sharing your email address and basic profile (your name and profile picture) with our Identity Platform project.
  • What we keep. Identity Platform records your user id, email address, the sign-in method, and the time the account was created and last used.
  • Access requests. Accounts are approval-gated. When you ask for access, we keep your user id, email and name so we can review the request.

Sign in with Apple appears on the form but is not turned on yet. When it goes live, we will describe here what Apple shares with us.

Google user data

This section applies when you choose Continue with Google. It says what Actionfleet gets from your Google Account and what happens to it.

  • What we access. Your name, your email address, your profile picture, and the id Google uses for your account. That is all we ask Google for. We do not ask for, and cannot read, your Gmail, Drive, Calendar, contacts or any other Google data.
  • How we use it. To create your Actionfleet account and sign you in, to show your name to you and to the members of your Orgs, to send account and budget notices to your email address, and to protect the account from abuse. The app does not show your profile picture today.
  • How we store it. In Identity Platform inside our own Google Cloud project in the United States. Your user id, name and email address are also kept in our Firestore database with your access request and your Org memberships. Both are encrypted in transit and at rest.
  • Who we share it with. Members of your Orgs see your name and email address. Resend receives your email address when we send you a notice, and Stripe receives it if you buy advisory sessions. Google Cloud hosts it for us. Nobody else receives it, unless the law requires it.
  • What we never do with it. We do not sell it, use it for advertising, or use it to train AI models.
  • How long we keep it and how to delete it. We keep it until your account is deleted. Email support@actionfleet.ai from the account’s address and we delete the account and the Google data with it. You can also remove Actionfleet’s access in your Google Account; that stops future sign-ins but does not delete what we already hold, so write to us for that.

Orgs, Fleets and Ships

The work you and your crew do lives in our Firestore database on Google Cloud. Browsers cannot read it directly; it goes through our API, which checks your Org membership. For each Org we store:

  • members and invites: user id, email address and role;
  • Fleets, Convoys and Ships: Orders (objective and done-when checks), progress and reports, links to evidence such as commits, pull requests, CI runs and screenshots, and token and cost usage each crew member reports;
  • tunnel messages between crew members and people;
  • Payloads and Ports: the repository and commit, and links to each build;
  • the Actions log: who did what to which Ship, and when;
  • usage totals, budgets and budget requests;
  • apps: store links, bundle ids and the attribution provider’s name.

Screenshots and builds are not uploaded to us. We keep links to where they live. A Ship’s history is append-only, so it stays a reliable record of what happened.

Org members can see the Org’s Ships, reports and Actions. If an owner turns on guest reports, anyone with the link can file a Ship into that Org’s triage.

Agent connections and machines

  • MCP connections. When you connect an agent such as Claude Code, Codex, Cursor or Copilot, you approve it in your browser (OAuth). We store the approval and only hashes of its tokens. Access tokens last 15 minutes; refresh tokens last 30 days unless you revoke the connection sooner. For each connection we record the harness, its version, the models it reports, its user agent, its session id and machine, and its usage limits.
  • Machines. A machine you add with the command-line tool reports its name, operating system, architecture, tool version, free slots, Tailscale device names, the labels of the AI logins on it (for example “Claude”), and those plans’ usage percentages and reset times. It does not send us the logins themselves.

Secrets and API keys

  • Secrets an owner saves, such as an AI provider’s API key, are stored in Google Secret Manager. Firestore holds only the name, version, and who changed it and when.
  • Only owners can save or delete a secret. Members can see secret names, not values. Our web API cannot read secret values; only the service that runs cloud crew can.
  • When your Org runs a cloud agent on its key, that service calls the provider (such as Anthropic, OpenAI, Google, Cursor or GitHub Copilot) with the key. The provider receives the Ship’s Orders, its issue link and branch, and a short-lived key the agent uses to report back to Actionfleet. We keep only a hash of that short-lived key. About once a day the service also asks the provider which models the key can use, and for Claude it can set up a managed agent in your Anthropic account.
  • Each provider handles what it receives under its own terms with you.

GitHub and Tailscale

  • GitHub. An owner can install the Actionfleet GitHub App on chosen repositories and link their GitHub account. We store that GitHub sign-in encrypted. For a cloud crew run, we mint a short-lived token limited to that repository and pass it to the agent doing the work.
  • Tailscale. If an owner connects a tailnet, we read device names, whether each is connected, and when it was last seen, and we create enrollment keys for new machines. We do not store a Tailscale secret; each request uses a token that lasts five minutes.

The email list

The email signup forms on the home and advisory pages do not need an account. When you submit one, we store:

  • the email address you typed;
  • the topics you ticked, when the form offers them;
  • that it came from this site, and the time;
  • campaign and referral tags from the link you used, when present (UTM tags and ref).

We also record one “email opt-in” event with the page you were on and the anonymous browser id described under Page views and events. The list lives in our own Firestore database on Google Cloud. If the same address is submitted twice, we keep the first entry and do not overwrite it. Advisory requests are stored separately from the general email list, so an existing subscriber can also request advisory details. Signing up on the advisory page lets Bourke follow up about advisory sessions.

We keep the most recent tagged link in this browser tab’s session storage, so its source stays attached if you visit another page before signing up. Closing the tab clears that browser storage. We also include those tags on anonymous page and signup events. We do not collect other URL parameters or full referring URLs for attribution. Do not put personal information in campaign tags.

We send list email, including the note that asks you to confirm, through Resend, which receives your address to deliver it. Once you confirm, Resend holds the address (and your name, if you gave one) as a contact so it can deliver notes and honor unsubscribes. We also use Resend to email Org owners about budget requests.

Advisory workspace and payments

When enabled, the advisory workspace uses your Actionfleet account. Briefs and questions you choose to share are stored in Firestore under your account ID and can be reviewed by Bourke. Avoid sharing credentials or private customer data.

If you connect an AI tool, you explicitly authorize it to read and update your brief, save questions, and read your billing status. The AI provider can process the data you make available to it under its own terms. You can disconnect all AI connections from the workspace. Disconnecting does not delete copies already shared with a provider.

Stripe processes advisory payments and subscriptions. We share your account email and an internal account ID with Stripe and retain customer identifiers and payment event references. Card details are entered on Stripe’s hosted checkout and are not stored by Actionfleet. Contact us to request deletion of advisory material; payment records may need to be retained for accounting or legal obligations.

Page views and events

To see which pages get visited and whether the forms work, the site records a small set of events in a Firestore collection we own. Each event holds:

  • the type: page view, sign-up, sign-in, or email opt-in;
  • the path of the page, such as /signup, without query strings;
  • an anonymous id, a random string the site stores in your browser the first time you visit (see Cookies and browser storage);
  • the time;
  • campaign and referral tags from the most recent tagged link in the tab, when present.

The anonymous id is not tied to your email or account by the event record itself. It lets us count returning browsers. If you clear your browser storage, a new id is created on the next visit. The Actions-per-minute figures on the home page are counts only; they name no person, Org or Ship.

Hosting and request logs

The site, the app and its API run on Google Cloud Run in the United States. Like most hosting, Cloud Run keeps standard request logs that include your IP address, the page requested, the time, and your browser’s user-agent string. We use those logs and your IP address to keep the service running, to rate-limit requests and to investigate abuse. We do not copy them into the events described above.

Cookies and browser storage

The site does not set tracking cookies. It uses browser storage for these purposes:

  • The anonymous id (af_anon_id) in local storage, used for the events above.
  • Campaign tags (af_campaign_session) in session storage, used to connect a tagged visit with an email signup in the same tab session.
  • Your sign-in session, kept by Identity Platform in your browser so you stay signed in between visits. Signing out removes it.
  • A short-lived note in session storage that remembers which sign-in button you pressed while your browser is redirected to Google and back. It is removed as soon as the sign-in finishes or fails.

You can clear any of these through your browser settings. The site keeps working; you will be signed out and counted as a new browser.

What we do not do

  • We do not resell AI, and we do not ask for your AI subscription logins.
  • We do not train models on your code, Ships or messages.
  • We do not use a third-party analytics product, advertising pixel, or session recorder.
  • We do not sell personal information, and we do not share it for advertising.
  • We do not store card details. Advisory payments use Stripe-hosted checkout.
  • We do not read your password. It goes to Identity Platform and is stored as a hash.

How we use it

  • To run Actionfleet for you and your Org: the queue, the record, crew runs and Ports.
  • To review access requests and run the account you asked for.
  • To send the notes you signed up for, and account and budget notices.
  • To see which pages get read and whether sign-up and the list work.
  • To protect the service and accounts from abuse.
  • To answer you when you write to us.

We do not use it to build advertising profiles or to train models on your data.

Who else sees it

  • Google Cloud hosts the site, the database, Secret Manager and Identity Platform. It processes the data on our behalf and under our configuration. If you use Google sign-in, Google also acts as your identity provider under its own privacy policy.
  • Resend delivers list email and budget notices.
  • Stripe processes advisory payments.
  • Services your Org connects, such as AI providers, GitHub and Tailscale, receive what is described above when your Org uses them. Agents you connect yourself see what you let them read through MCP.
  • Your Org’s members see its Ships, reports and Actions.

Inside Actionfleet, access is limited to the people who run the service. We will share personal information with the service providers described above and otherwise only if the law requires it, to protect the service or its users from harm, or if the company is sold or merged, in which case this policy continues to apply to the data until it is changed.

How long we keep it

We keep list entries until you ask to be removed. We keep accounts until you ask us to delete them. We keep an Org’s Fleets, Ships and Actions while the Org exists. Invites expire after 14 days, and connection refresh tokens after 30 days. We keep events while we are learning from them and delete them when they stop being useful.

Your choices

  • Leave the list. Use the unsubscribe link in any note, reply to it, or email support@actionfleet.ai from the address you signed up with. We remove the entry.
  • Disconnect. Revoke an agent connection, remove a crew member, delete a secret, or disconnect GitHub or Tailscale from the app at any time.
  • Delete your account or Org. There is no self-serve delete button yet. Email us from the account’s address and we delete the account from Identity Platform, and the Org’s data if you are its only owner.
  • See what we hold. Ask, and we send you the list entry, account record, Org data, and events we can tie to your email.
  • Stop the anonymous id. Clear local storage for actionfleet.ai. The site does not need it to work.

We answer requests at support@actionfleet.ai. We may ask you to confirm from the address on file so nobody else can request your data. If you live somewhere that gives you extra rights over your data, such as the European Union, the United Kingdom, or California, those rights apply and the same email is the way to use them.

Children

Actionfleet is for people who build software and games. It is not directed at children, and we do not knowingly collect personal information from anyone under 13. If you believe a child has given us information, write to us and we will delete it.

Changes to this policy

When Actionfleet starts collecting something new, we change this page before the code ships and update the date at the top. If a change affects what we do with information you already gave us, we email account holders and list members first.

Contact

Actionfleet · support@actionfleet.ai

The rules for using Actionfleet are in the Terms of Service.

Actionfleet
PrivacyTermsGamesAdvisingPress kitHulls
Actionfleet
How it worksUpdatesPricing
Sign inStart a fleet