Legal

Privacy Policy

Effective · Principled Labs, Inc., doing business as orgasm.dev

orgasm.dev measures something intimate about a person. This policy says what information we handle, why, for how long, who else sees it and what you can do about it: for visitors, for developers, and for the people on camera in the products they build.

Part I · Who We Are

1. Scope and who we are

Principled Labs, Inc., doing business as orgasm.dev, is responsible for your information; this policy covers the website and its documentation, the dashboard, the API and its sandbox, the device SDK, the MCP servers and our emails.

This Privacy Policy explains how Principled Labs, Inc., doing business as orgasm.dev (“orgasm.dev”, “we”, “us”) handles personal information in connection with the orgasm.dev website and its documentation, the developer dashboard, the orgasm.dev API and its sandbox, the device SDK, the MCP servers at orgasm.dev/mcp and orgasm.dev/api/mcp, and the email we exchange with you (together, the “Service”).

It concerns four groups of people:

  • Visitors to the website and its documentation.
  • Developers, who sign in to the dashboard and build products with the API, and the organisations they build for.
  • Creators, who sign in to the Studio to put the HUD on their own streams.
  • The people on camera, who use a developer’s product that sends their camera to orgasm.dev for analysis.

What developers must do before anyone is analysed (adults only, consent first, never covertly) is set out in our Acceptable use rules. This policy describes what we do with the information.

2. Our two roles

For visitors and developers we decide how information is used; for the people on camera we process it on behalf of the developer whose product they use, as its service provider.

For visitors, developers and creators, orgasm.dev decides how their information is used: we are the business responsible for it (under the GDPR, the controller).

For the people on camera, the developer whose product they use decides to analyse them, what the results are for and how long they are kept. We process that information on the developer’s behalf and on its instructions, as its service provider (under the GDPR, its processor). The developer’s own privacy notice explains what it does, and the developer is responsible for telling the person, verifying that they are an adult and obtaining their consent before anything is analysed.

The one use we make of that information on our own account is the de-identified, aggregated statistics described in section 10.

3. What orgasm.dev does

An API that reads, from a person’s front camera and microphone, how close they are to orgasm and when one begins, peaks and ends; it is not a medical device, and it never identifies anyone.

A developer’s product creates a session through the API. The person’s own device then verifies a confidential computing enclave and sends it the front camera, usually the microphone and, where the product offers it, a second view of the body. The enclave measures what it sees and hears and returns events:

  • the probability of orgasm within the next 30 and 60 seconds, about once a second;
  • the onset, peak and end of each orgasm, a few seconds after they happen;
  • in a live session, about once a second, how often the person vocalises and how loudly, their pelvic floor contractions and movement where a view of the body is sent, how far each region of the face has moved from their own resting state, and the face’s movement as expression channels and the turn of the head, which a display can draw on a generic face that is not theirs (and, for such a display, about five times a second);
  • whether the face reads as a woman’s or a man’s, which decides the detectors that apply;
  • notices about the input, such as a face out of view.

orgasm.dev is not a healthcare provider, and its events are not a diagnosis or a treatment. It does not recognise or identify people: it never compares a face with other faces, and never keeps one to recognise it later.

Part II · Developers, Creators and Visitors

4. What we collect from developers and visitors

Your work email address and sign-in details, the company it belongs to, your role in it and that company’s credits and charges, what you tell us when you set up your account, your keys (as hashes), webhook endpoints and usage, technical information about your requests including your IP address, and what you write to us; never your name or photo.

From developers and visitors we collect:

  • Sign-in details. When you sign in with Google, or with Apple where offered, the provider gives us an identifier for your account and your email address, and tells us whether the address is a private relay address, such as Apple’s Hide My Email. We ask for nothing else: not your name, your photo or your contacts. We record when you first and last signed in, how many times, and the part of the site you first signed in from. Only an address at a company’s own domain can sign in to the dashboard: when the address is a personal mailbox or a private relay, the sign-in is refused and we keep nothing from it. Creators sign in to the Studio with any address (section 7).
  • Your company. The domain of your email address, and your place in its company on orgasm.dev: whether you are a member, asked to join or were removed, your role (Admin, Developer or Billing), who invited or let you in, and when. Members share the company’s credits, keys and webhook endpoints and, in the dashboard, see each other’s email addresses, roles, and when each joined and last signed in. Admins also see the invitations (the address, the role and who made it), the requests to join (the address and when), and a log of changes to the company’s keys, webhook endpoints, GPUs and members, with who made each and when. We email these steps: an invitation to the address invited, with who invited it; a request to join to the company’s Admins; and the answer to the person who asked. For the company we keep its credits: each package it received, with its price per credit, the statement of work or invoice it was sold under and when it was paid, posted and expires; and each charge against them, with the session, what was counted and when.
  • What you tell us when you set up your account. What you are building first, and that you accept our Acceptable use rules, with the version you accepted and when.
  • Your account. The API keys you create: of each secret key we keep only a one-way hash and a shortened form to show you, with the name you give it, who made it, and when it was created, last used, rolled and revoked. The webhook endpoints you register: their address, description, the events they receive and the secret we sign deliveries with. Daily counts of your sessions and their duration, which the dashboard shows as usage.
  • What you send through the API. The settings, metadata and references of the sessions you create. Metadata and references are yours to define, and must not contain anyone’s name or contact details. In the sandbox, the scenarios you choose and, if you try an upload, the size of the file: its content is counted and discarded, never stored.
  • Technical information. Your IP address, your browser or client, and the pages and API addresses you request, which reach our servers and our hosting providers with each request.
  • Correspondence. What you send us by email, for example to legal@orgasm.dev, and our replies.

We do not run analytics or advertising tools on the website, and we do not buy information about you from anyone.

5. How we use it

To run your account and the API, email your company’s invitations and requests to join, keep and charge your company’s credits, keep the Service secure, keep a record that you accepted our Acceptable use rules and enforce them, point you to the right guides, tell you what changes and meet the law; never to sell it or advertise to you.

We use the information about developers and visitors to:

  • sign you in and keep you signed in;
  • let your company’s Admins choose who joins it and with what role, email the invitations, the requests to join and their answers, and show its members who changed its keys, endpoints and members;
  • authenticate API requests made with your keys, deliver events to your webhook endpoints and show you your usage;
  • keep your company’s credits, charge them for what its live sessions count, tell you when they run low, and keep the record of purchases and charges our accounts and the law require;
  • let our sales team contact you about buying credits for your company;
  • keep the Service secure: we use IP addresses, in memory only, to limit how many requests one address can make, and our logs to investigate faults and abuse;
  • keep a record that you accepted our Acceptable use rules, ask you to accept them again when they change, and enforce them, for example by revoking keys used to analyse someone without their consent;
  • point you to the guides and examples for what you are building, and understand who builds with orgasm.dev;
  • write to you about your account, its security, and changes to the Service or to this policy;
  • comply with the law and protect our rights and the safety of others.

We do not sell your personal information, share it for cross-context behavioural advertising, or use it to advertise to you.

Information from Google. When you sign in with Google we receive only your Google Account’s identifier and email address. We use them only to sign you in, keep your account and write to you about it. We do not use them for advertising, we do not sell them, and we transfer them to no one except the service providers in section 11, as needed to run the Service, or as the law requires. Our use of information received from Google APIs adheres to the Google API Services User Data Policy.

6. Cookies and similar technologies

Essential cookies of our own for signing in to the dashboard and the Studio, your display preferences in your browser, and nothing that tracks you.

The dashboard and the Studio set cookies of our own, all essential to signing in: one holds a sign-in in progress for ten minutes; one keeps you signed in to the dashboard, and one to the Studio, each for up to 90 days or until you sign out. None is used for anything else.

The website keeps your display preferences, such as the light or dark theme, in your browser’s storage. It sets no analytics or advertising cookies and loads no third-party trackers. One page, the API reference, loads its viewer from jsDelivr, a public content delivery network, and its fonts from Scalar, the viewer’s maker; like any site your browser loads files from, they receive your IP address and browser details.

Because we neither sell personal information nor share it for advertising, there is nothing to opt out of; we nonetheless treat a Global Privacy Control signal as a request to opt out of both.

7. Creators

Creators sign in at the Studio with any address; we keep their sign-in, their overlay’s address, a hash of its link’s key and what it shows, and anyone with the link sees what it shows.

Creators sign in to the Studio to put the HUD on their own streams. Any address can sign in there, a personal mailbox or a private relay included, and a creator’s account is kept apart from any company’s. From creators we keep:

  • Sign-in details, as for developers: your account’s identifier at the provider, your email address and whether it is a private relay, and when and how many times you signed in.
  • Your overlay: its address, a code of eight letters; a one-way hash of its link’s key, and when you made the link; and which parts of the HUD it shows. The face, a generic face that moves as the person’s does, is shown only if you turn it on.
  • Your age verification: when you verified your age, once verification is available. Until then we record that you have not.

A rehearsal is a scripted session: nobody is on camera, nothing about you is analysed, and it is kept only in memory while it plays. Checking your camera in the Studio happens in your browser alone, and nothing from it is sent. Anyone with your overlay link sees what your overlay shows while a session plays on it, so share it only where you mean to; making a new link stops the old one at once.

Your own live sessions open once age verification is available. Before they do, this policy will say how they are handled, and you will be asked to agree to it.

Part III · The People on Camera

8. What is processed about the people on camera

The video and sound from the person’s device, analysed in an enclave and never stored; the events the analysis produces; the session’s settings, and the metadata and pseudonymous references the developer attaches; and the device’s IP address. We do not learn who the person is.

When a developer’s product starts a session, we process:

  • Video and sound. The person’s front camera and, where the product sends them, the microphone and a view of the body. They are analysed inside the enclave and never stored (section 9).
  • Events. What the analysis returns: the probability of orgasm within the next 30 and 60 seconds; the onset, peak and end of each orgasm; in a live session, how often the person vocalises and how loudly, their pelvic floor contractions and movement where a view of the body is sent, how far each region of their face has moved from their own resting state, and the face’s movement as expression channels and the turn of the head; whether the face reads as a woman’s or a man’s, used only to choose the detectors; and notices about the input, such as a face out of view or more than one face. The face’s channels sent about five times a second for a display that draws them are not events: they are passed to that display and never stored. The read of the face is not a finding about the person’s sex or gender, and must not be used as one.
  • Session details. The settings the developer chose (the inputs, the longest the session may run, how long its events are kept), the metadata and references it attached, the reports of stimulation its product sends (whether a device is on, its level, its outputs’ levels and its mode, which the session’s readers receive as its events), a hash of a random key that the person’s device generated for the session, and what we counted in it to charge the developer: how many orgasms, ruins and edges, and when. The key itself never reaches us.
  • The device’s network address. The device’s IP address reaches the enclave when the device connects to it, and our servers when the device fetches its connection details and its events. Where the device’s network needs one, the connection passes through our relay (a TURN server), which carries the media encrypted and cannot read it.

We do not receive the person’s name, email address or account with the developer, and we do not try to find out who they are. A developer may attach to a session a pseudonymous reference of its own for the person, so that it can see what each of its users cost: an identifier we cannot link to them, and never an email address, phone number or account name, which we refuse where we can recognise one. Developers must not put such details in a reference or in a session’s metadata. A session’s references and metadata are erased when the developer purges the session, and in any case 13 months after it; what it was charged stays, without them.

9. How video and sound are handled

They are decrypted only inside a confidential computing enclave that the person’s device verifies first, turned into measurements there and discarded, never written to disk, logged or seen by anyone.

Video and sound are decrypted and analysed only inside a confidential virtual machine on Google Cloud (Confidential Space). The processor encrypts its memory and its graphics processor runs in confidential computing mode, so neither the cloud provider nor our operators can read what it analyses. It runs a signed image whose code that reads video and sound is public, with our detectors pinned in it by hash, and Google’s attestation service signs a statement naming that exact image and hardware, which anyone can check against our published attestation policy.

Before sending anything, the person’s device checks that statement and encrypts the session so that only that enclave can open it. Inside, frames and sound are turned into measurements and discarded as they are analysed: they are never written to disk, never logged and never leave the enclave, which writes no logs outside itself. Only the events leave it, carried by our relay service to our servers. Trust explains each step and where the guarantee stops.

The guarantee holds when the person’s own device sends its camera straight to the enclave, as our SDK does. If a developer routes video through its own servers instead, that developer can see it, and its own notice must say so.

A developer may also run the enclave on its own GPUs (your own GPUs). The confidential virtual machine is then in that developer’s Google Cloud project instead of ours, running the same signed image under the same checks: the device verifies it the same way before sending anything, the developer cannot read what it analyses, and it keeps no record of a session. Its events still come to our servers, each signed by the enclave, and are handled as this policy says.

Recorded video files are not analysed yet: with a live key the API declines them, and the sandbox, which plays only scripted sessions, accepts a test file only to count its size.

10. How we use information about the people on camera

To provide the analysis to the developer who asked for it, on its instructions, and to make de-identified, aggregated statistics that improve our detectors; never to identify, track or advertise, and never sold.

On the developer’s behalf we use this information to run its sessions: to analyse them; to send the events to the developer as they happen, as server-sent events, over a WebSocket or to its webhook endpoints; to keep them readable for the period it chose; to end sessions that run too long or never start; to count each session, its duration and its orgasms, ruins and edges towards the developer’s usage and charges; and to keep the Service secure.

On our own account, we may combine events from many sessions into de-identified, aggregated statistics, for example how often each detector calls an onset at each level of imminence, and use them to improve our detectors. These statistics contain no video, sound or images, no session, device or developer identifiers, and nothing that relates to any one person. We keep and use them only in that form, we do not attempt to re-identify anyone from them, and we require anyone we share them with to do the same.

We never use information about the people on camera to identify, recognise, track or profile them, for advertising, or for decisions about them, and we never sell it. It goes only to the developer who started the session and, as section 11 describes, to our service providers. Developers must tell the people on camera about this processing, the aggregated statistics included.

Part IV · Sharing, Keeping and Protecting

11. How we share information

With our service providers under contract (Google Cloud, Cloudflare, and Customer.io for our email), the sign-in provider you choose, an affiliate that runs our infrastructure, the developer who started a session, and the authorities when the law requires.

We share personal information only as follows:

  • Google Cloud hosts the API, its database, the confidential computing enclaves, our relays and our logs, and carries our email through Google Workspace (Google Cloud Privacy Notice).
  • Cloudflare delivers the website, passes requests on to the API and protects both from attacks (Cloudflare’s privacy policy).
  • Customer.io sends the Service’s email for us: invitations to join a company, requests to join to its Admins, and their answers. It holds the addresses it emails and, for developers, what we tell it when you sign in: your account’s identifier, your email address, your sign-in provider, whether the address is a private relay, when and how often you signed in, and the part of the site you first signed in from. Our emails carry no tracking pixels or tracked links (Customer.io’s privacy policy).
  • Google and Apple sign you in when you choose them, and learn that you signed in to orgasm.dev, under their own policies (Google, Apple).
  • jsDelivr and Scalar serve the API reference page’s viewer and fonts to your browser when you open it (section 6).
  • Our affiliate, a company under common control with us, provides engineering and infrastructure services, among them the relays that carry readings from the enclaves and media from devices. It acts for us and is bound by this policy.
  • The developer who started a session receives its events and details at the endpoints it chooses; from there, the developer’s own privacy notice applies.
  • Legal and safety. We disclose information when the law requires it, or when we believe in good faith that it is needed to protect someone from harm, to investigate misuse such as analysis without consent or of a minor, or to protect our rights.
  • Business transfers. If we are involved in a merger, an acquisition or a sale of assets, information may pass to the successor, which remains bound by the promises in this policy for the information it receives.

Our service providers and our affiliate may use the information only to provide their services to us. We do not sell personal information, and we do not share it for cross-context behavioural advertising.

12. How long we keep information and how we protect it

Developer accounts until you ask us to delete them; a company’s credits and charges as long as our accounting records require, and its team as long as it has an account; logs for up to 30 days; a session’s events for the period the developer chose, 30 days by default and 90 at most, and its references 13 months; video and sound never.

  • Your developer account, what you told us when you set it up, its keys, webhook endpoints and usage: for as long as the account exists. A revoked key’s record stays with the account and can no longer be used. When you ask us to delete the account, we delete it within 45 days, except for anything the law requires us to keep.
  • Your creator account, its sign-in details and your overlay: for as long as the account exists. When you ask us to delete it, we delete it within 45 days.
  • Logs of requests, which include IP addresses, and of our relay’s connections: at most 30 days.
  • Correspondence: for as long as we need it to answer you and to keep a record of the requests we received.
  • Video and sound from the people on camera: never stored.
  • Events and session details: while the session runs and, after it ends, for the period the developer chose, 30 days by default and at most 90, or not at all; then we delete them. A developer can delete a session at any time, which deletes it at once. What a developer has already received is in its own systems, under its own policy. A session’s charges (its identifier, what was counted and when, and nothing about the person) stay with the developer’s company’s credits.
  • A session’s references and metadata, kept with its charges so the developer can see what it cost: 13 months after the session, or until the developer purges it; then we erase them.
  • Your company’s credits, its packages and every charge against them: for as long as the company has an account, then for as long as the law requires us to keep accounting records.
  • Your company’s team (its members and their roles, invitations, requests to join and the log of changes): for as long as the company has an account. A member who leaves or is removed stays in the log of what they changed.
  • De-identified, aggregated statistics: for as long as they are useful, since they relate to no one.
  • Sandbox sessions, made with test keys: scripted, and held in memory only until the service next restarts.

We protect information with encryption in transit and at rest, keep secret keys only as hashes, limit our staff’s access to what their work needs, and analyse media only inside verified enclaves. No system is perfectly secure; if a breach affects your information, we will tell you as the law requires.

13. Sensitive, biometric and health-related information

The events describe a person’s sexual response and come from their face; we treat them as sensitive, never sell them, keep any face geometry only in the enclave’s memory while a frame is analysed, and rely on developers for the consent the law requires.

Information about a person’s sexual response is sensitive personal information under U.S. state privacy laws and special-category data under the GDPR and the UK GDPR, and it may be consumer health data under the laws of Washington, Nevada and Connecticut. The analysis of a face may also involve its geometry, which some laws, among them those of Illinois, Texas and Washington, treat as biometric information. We treat all of it as sensitive.

Our retention schedule for biometric information. Any measurement of face geometry is computed only in the enclave’s memory, from frames that are discarded as they are analysed, and is destroyed as soon as its frame has been analysed. It is never stored, logged or sent out of the enclave; only the events leave it. The events that describe the face’s expression say how far each region and each expression channel has moved from that person’s own resting state, and how far the head has turned, not the face’s shape: a display draws them on a generic face that is no one’s, so they show how the face moves and not what it looks like, and they are never used to identify anyone. The same channels sent about five times a second for such a display are passed to it and never stored. Every other piece of information about the people on camera is permanently destroyed when its period in section 12 ends, and in no case later than 90 days after the session ends. We never sell, lease or trade biometric information or otherwise profit from it, never use it to identify anyone, and disclose it to no one.

Developers are responsible for the notices and consents these laws require before a session starts: among them, a written release where Illinois law requires one, and consent to the collection and sharing of consumer health data where Washington or Nevada law requires it. We process the information only as section 10 describes.

Part V · Your Rights and Our Notices

14. Your choices and rights

Developers can see, copy, correct and delete their information by writing to us; the people on camera should ask the developer first and can always write to us; the rights of U.S. state, EEA and UK residents explained.

Developers and visitors. You can revoke keys and delete webhook endpoints in the dashboard, and sign out at any time. To see or get a copy of your information, correct it, or delete your account, write to legal@orgasm.dev from the address you sign in with.

The people on camera. We process your session for the developer whose product you used, so please make requests about it to that developer, who can delete a session at once; we will help it respond. If you cannot reach the developer, or believe orgasm.dev was used on you without your consent, write to legal@orgasm.dev. We will act on what we can verify, which may include ending sessions and revoking the developer’s access. We hold no name or contact details for you, so we may need details of the session, such as when it ran and in which product, to find it.

U.S. state privacy rights. Depending on where you live, for example in California, Colorado, Connecticut, Oregon, Texas, Virginia or Washington, you may have the right to know what personal information we hold about you and how we use and disclose it; to get a copy of it; to correct it; to delete it; to opt out of its sale, of sharing for cross-context behavioural advertising and of targeted advertising, none of which we do; to limit the use of sensitive personal information, which we use only to provide the Service; and not to be treated differently for using these rights. You may use an authorised agent. We verify a request through the email address on the account or, for the people on camera, with the developer, and answer within 45 days, which the law allows us to extend once. If we decline a request, you can appeal by replying to our answer; if the appeal is refused, you can contact your state’s attorney general.

California. In the past 12 months we have collected these categories of personal information about developers and visitors: identifiers (an email address, an account identifier, an IP address); account credentials (API keys, kept only as hashes); professional information (what you are building, the company your address belongs to and your role in it); commercial information (your company’s credit purchases and charges); internet or network activity (requests and logs); and the contents of messages sent to us. The sources are you, your sign-in provider and your use of the Service; the purposes are those in section 5; the recipients are those in section 11. For the people on camera, we act as the developer’s service provider. We have not sold personal information or shared it for cross-context behavioural advertising, and we have no actual knowledge of doing so for anyone under 16.

The EEA and the UK. If the GDPR or the UK GDPR applies to you, we rely on performing our contract with you for developer accounts; on our legitimate interests in running, securing and improving the Service and in preventing its misuse, for logs and security; and on our legal obligations where the law requires us to act. For the people on camera, the developer’s legal basis applies, which for this kind of information is normally the person’s explicit consent, and the developer’s notice must also cover the aggregated statistics in section 10. You have the rights of access, rectification, erasure, restriction, objection and portability, and you may withdraw consent at any time. You may also complain to your data protection supervisory authority.

15. International transfer

We are in the United States, and information is processed there.

We are in the United States and run the Service from there. Our servers, the enclaves and our database are in the United States, so information is processed there, including the analysis of the people on camera; Cloudflare handles each request at the data centre nearest to it before passing it on. If you use the Service from elsewhere, your information is transferred to the United States, whose laws may differ from those where you live.

16. Adults only

No one under 18 may ever be analysed, and we do not knowingly process information about minors.

orgasm.dev is for products made for adults. The Service must never be used on anyone under 18, or under the age of majority where they live, and developers must verify the age of everyone on camera before a session starts; a self-declared birthday is not verification. We do not knowingly collect or process information about minors. If you believe a minor has been analysed, write to legal@orgasm.dev: we will end the sessions we can find and delete their data.

17. Changes to this policy

The effective date shows the current version; we tell developers before a material change takes effect.

We update this policy when the Service changes; the effective date at the top shows when it last did. Before a material change takes effect we tell developers by email or in the dashboard, early enough for them to update their own notices to the people on camera.

18. How to contact us

legal@orgasm.dev for questions, requests and reports of misuse, or Principled Labs, Inc., 1550 Tiburon Blvd, Suite G-207, Tiburon, CA 94920.

Questions, requests and reports of misuse, including analysis without someone’s consent or of a minor: legal@orgasm.dev. By post:

Principled Labs, Inc.
1550 Tiburon Blvd

Suite G-207

Tiburon, CA 94920

United States