Who we are, what we hold, and how to make it stop
Last updated: 17 September 2026
A game studio contacted you from its own mailbox using Pigeoneer. We provide the contact database and process campaign messages on the studio's behalf. If a message about a video game arrived in your inbox, the studio in the From line sent it, and we are the ones holding the information that helped them find you. This notice explains what we hold about you and how to exercise your rights.
You are most likely here because you asked that studio how they got your address, because you went looking for whoever is behind the tool, because you found the link in the footer of our website, or because you published something about a game and want to know what we do with that. This page covers all of it.
It explains who we are, what we hold about you, and how to make it stop.
Make us stop
To stop the campaign that wrote to you: reply to that email and say stop. It goes back to the studio's mailbox, and we see it there. The word "stop" is enough, and you can say it in your own words instead.
The system records it for that campaign without waiting on a person, the studio cannot lift it, and no later import or re-scoring brings you back. No account, no form, no reason, no login. One limit: recognising a stop written in your own words is automated, and an unusual phrasing can slip past it. If anything reaches you after you asked, tell us at the address below and a person records it.
What a stop covers follows what you said, not where you sent it. Stop, on its own, read in the context of your reply, ends that campaign to you. Name the game and it ends every campaign for that game, from any studio. Name the studio and it ends every campaign of that studio. Say you want nothing through Pigeoneer, or no marketing at all, and it ends every studio's campaign and every other marketing use we make of your record: no more scoring, no more research notes, no more showing your profile to a studio. If it is not clear which you meant, we read it the wider way. Only the single-campaign stop is applied by the software today; the three wider ones are applied by a person, across every campaign they reach, and entered on a written list we check before every import, every scoring pass and every send.
To stop hearing from anyone through Pigeoneer: email [email protected] and say so, or say it in your reply to the studio. We take you out for every studio we work with, and we will confirm when it is finished. You do not need to use any particular wording, and there is only one address to remember.
If you want us gone completely, ask for deletion and we will remove your record: your profile and the notes and scores we generated about you, at once; our copies of anything you wrote, and any summary we made of coverage you published, starting at once, under a standing instruction in the terms every studio accepts. What is left belongs to the studio's campaign (its own copy of your reply in its mailbox, and what its software made of your reply, unless you ask us to delete that too), so we pass that part of your request to the studio within three business days and tell you we have. One thing stays if you have also told us to stop: the minimum needed to keep honouring that: your address and handle, and the fact, date and scope of what you asked, nothing else. We keep it because without it nothing would stop your address being found again from the same public page it came from and you being written to again. If you ask for deletion without ever having asked us to stop, we keep no opt-out record either, and we will tell you what that means before we do it.
We do not ask you for anything else, and we will not try to talk you out of it.
The email was written by an AI
The message you received was drafted by an AI system, using a large language model, from public information about your channel. The game studio decides who is contacted and chooses how messages are released: some studios read and approve every message before it goes, and some let a queue send on a timer, in which case no person read that particular message before it reached you.
If you reply to the studio, the answer you get back may also be AI drafted, released under the same settings. It still comes from the studio's mailbox, not from us. If you want a person to answer you, say so in your reply; the studio decides how it answers, and if you write to us instead we pass your message to them.
Nothing about your reply feeds an automated decision that has any legal or similarly significant effect on you. Nobody is scored for creditworthiness, employment, or anything of that kind.
Who we are
Pigeoneer is a tool that helps small game studios find and contact creators and journalists who might be interested in their game.
Pigeoneer LLC, a California limited liability company, is the controller of the shared database described on this page: your public profile, the research notes and scores we generate about your channel, our own usage records, and any opt-out record. The studio that wrote to you is the controller of everything that belongs to its campaign: your reply, what our software makes of that reply, and any summary of coverage you published about its game; we hold those for the studio, as its processor. Ask at the email below if you need our registered postal address or our agent for service.
Contact for anything on this page, including all of the rights below: [email protected].
The game studio that emailed you is a separate controller for its own campaign, and the message came from its own mailbox rather than from us. If you want to deal with them directly, they are in the From line. If you want to deal with the tool and the database behind it, that is us, at [email protected].
Representatives in the EU and the UK. We are appointing a representative in the European Union under Article 27 GDPR and one in the United Kingdom under Article 27 UK GDPR, and we will name them here with their contact details as soon as they are in place. Until then, every right on this page is exercised directly with us at [email protected], and you may also complain directly to your own supervisory authority without going through us.
One split you should know about. Two different things happen to two different kinds of data. Everything we found about you from public sources, and the notes and scores we generated from it, we are responsible for: we decide it, we hold it, and you deal with us. Anything you wrote back to a studio, what our software makes of it, and any summary of coverage you published about that studio's game, is the studio's: we hold it on their behalf, in their campaign, for as long as they keep their account with us, so they can carry on the conversation and keep the history of their own campaign; the "How long we keep it" section below says what that means. A copy of your reply also sits in the studio's own mailbox, which is theirs and outside our control. If you ask us to delete your reply, we start deleting our own copies at once, pass the request to that studio within three business days and tell you we have, and finish within thirty days, with backup copies going at the next rotation. If you ask us to delete you from our database, that is entirely our decision and we will just do it.
What we hold about you
The public part. Collected from your public channel and profile pages:
- your handle, channel or display name, and channel URL
- your country and the main language you publish in
- your public subscriber, follower and view counts
- the contact email you publish on your channel or site
- your public social handles and website
- your channel description or bio text
- the titles of your recent videos or streams
Information we generate. We also hold things we generated about you. They did not come from any public source. They came from us:
- a fit score, a number estimating how well your channel matches a particular game
- an AI written rationale for that score, in prose, explaining why you were considered a match
- notes on your content, also AI written, summarising what you cover and what an opening line to you might mention
- if you replied to a studio: a stored snippet of your reply, roughly the first 400 characters, and in some cases the full text of the message, held in that studio's campaign records and not on your shared record
- an AI assessment of your reply, its apparent sentiment and its apparent intent
A language model wrote that material. It is inference, not fact, and it can be wrong about you. It can misread your tone, misdescribe your channel, or draw a conclusion you would not recognise. You have the right to see it and the right to have it corrected or deleted. Email [email protected] and ask.
One more record, which is ours and is not about your channel. Each time our software runs a model about you, we log which model, what for, how many tokens and what it cost, tied to the campaign and, today, to your handle. That is our own business record, kept to run and cost the platform, for as long as we operate it. We strip your handle from it when the studio's account closes, or when you ask us to delete you, whichever is first.
How the scoring actually works, and what it does to you. A language model reads your public channel data: the games and genres in your recent titles, your bio, your language and country, your audience size, and whether your channel has covered comparable games. It produces a number from that, plus a short written reason. Nothing in the score comes from anything you have not published. The score is then used for one thing: to rank who a studio's campaign contacts first, and to cut off the list somewhere. A high score means you may get one email. A low score means you get nothing, and you will not know it happened. The score is specific to one game, so a low score against a puzzle game says nothing about the next studio's roguelike, and we re-run it for each campaign. No score of ours decides anything else about you: no money, no access, no platform standing, no opportunity you had applied for. You can ask us for your current scores and rationales for any campaign, and you can tell us they are wrong.
We do not go looking for special category data, the sensitive kinds such as health, politics, religion or sexual orientation, and nothing detects it automatically. If something like that ends up in a stored reply because you wrote it, we do not use it to score or target you, and if you tell us we will remove that text.
If you publish something about a game
This part is separate from the outreach, and it applies whether or not we ever wrote to you.
When a video or stream is published about a game a studio is running a campaign for, we look at it. We take the public information the platform gives us about the piece, including its title, publication date, view count and comment count, and a transcript: for a YouTube video, the caption track published with it where there is one; otherwise, and for every Twitch stream, one we make ourselves by downloading the audio and running speech-to-text on our own machine. We use the transcript to produce a short written summary for the studio: how the piece received the game, and the moments in it worth their attention, such as a player getting stuck or running into a bug.
Three things worth saying plainly about that:
- The studio never sees the transcript. Their reports contain summaries, not transcripts. The underlying text is not available to them through the product, and they are contractually forbidden from trying to reconstruct or extract it.
- It is meant to run for a set number of weeks. A studio's campaign comes with a tracking window, and the studio's terms limit new coverage discovery to it. The software does not yet stop discovery at the end of that window: today it runs for as long as the studio's project is active.
- This happens whether or not you were contacted. If you covered the game entirely off your own bat, having never heard from us or from the studio, your video is still something we may look at, because we are looking for coverage of the game rather than for coverage by particular people.
- We tell the studio which of the people they wrote to went on to publish. That is a list of who was approached and who published, and nothing more. We tell them in writing that publication after a message is a coincidence in timing and not proof their message caused anything, and that a summary that does not mention a free key is not evidence that anything was hidden from you.
The caption or audio file lives in a temporary folder on our machine for the length of the transcription and is deleted when it finishes, whether or not it succeeded; the transcript text is held in memory for the summarising call and is not stored by our application. That text is sent to the AI provider named on the subprocessor page, which writes the summary and, under its commercial terms, keeps what we sent it for up to thirty days before deleting it; the audio and video never leave our machine. What we keep is the summary, with the transcript's source, length and language. We intend to move the summarising step to our own machine as well, and will update this page when that is done. A transcript can contain things you said about yourself that are sensitive. Nothing detects that automatically; we do not use it to score or target you, and if you tell us we remove that passage from the summary and from our copies.
If you would rather we did not do this in relation to you, email [email protected] and say so: we stop summarising your content, and if you also ask, we delete the summaries we hold.
Where we got it
- the public YouTube Data API and the public Twitch API, and the caption tracks and audio of published videos and streams, fetched from YouTube and Twitch
- public web pages, including your channel About page, your own website, and public third party stats pages
- transcripts and public statistics for videos and streams published about a game, as described above
- your own reply, if you sent us one
We do not buy lists, and we did not get your details from another marketing company.
Why we hold it, and on what legal basis
For the shared database, your public profile and what we generate from it, our legal basis under the UK and EU GDPR is legitimate interests, Article 6(1)(f). For your reply and anything made from it, and for coverage summaries, the studio is the controller and relies on its own basis, which it should tell you about. For an opt-out record, our basis is the legal obligation to honour your objection, Article 6(1)(c) with Article 21(3). Concretely, our interests for the shared database are:
- letting a small game studio reach creators who plausibly care about their game, instead of blasting everyone
- keeping the outreach relevant, so you get one email about a game in your genre rather than five about games in nobody's genre
- keeping a record of who has been contacted and who has said no, so you are not contacted twice
- reading published coverage of a game so the studio can understand how it landed and find problems in their own product
We have written that balance down, purpose by purpose, and we keep the assessment on file. Ask us for a summary at [email protected] and we will send you one. You may disagree with our conclusion, and if you do, your objection wins. See the next section.
For direct marketing, once you object we cannot weigh our interests against yours: under Article 21(2) and (3) you can object at any time and we must stop. That is the "make us stop" section above, and if what you want is to be out of it entirely rather than out of one studio's campaign, the second and third routes there are the ones to use.
Who sees it
- A studio choosing whom to contact. Before any email is sent, a studio running a campaign sees, for the creators our scoring matched to its game, your public profile, the score and the rationale. That happens whether or not it then contacts you.
- The game studio running a campaign that contacted you, or whose game you covered. A studio that contacted you sees your profile, the score and rationale, and your reply. A studio whose game you published about sees the summary described above, whether or not it contacted you, but never the transcript. Their terms with us bind them: the profiles, scores and research we show them are for running their campaign with us and may not be sold, passed on or used for anything else; the coverage summaries may be used only to run and judge their campaign and improve their game, may not be republished, and may not be used to harass, coerce or retaliate against you or to make keys or access depend on how you speak about the game. Your reply stays with them as their own correspondence for as long as they keep their account; the next section says how long that is.
- Our service providers, listed at pigeoneer.app/subprocessors. That list includes the AI providers whose models draft the emails and write the notes and summaries described above.
We do not share your data with advertising networks. We do not put tracking pixels in the emails.
We are a United States company and our systems are operated from the United States (California). Your public profile and what we generate from it we collect directly from public sources, so that part is not "transferred" out of the UK or the EEA by anyone else; it is gathered by us, where we are. Your reply is different: it arrives in the studio's mailbox and we read it from there, so for a studio in the EEA it reaches us under the European Commission's standard contractual clauses between that studio and us, which you can ask us about. Our service providers, listed at pigeoneer.app/subprocessors, are in the United States. Each of them has the European Commission's standard contractual clauses in place with us, and two of the three are additionally certified under the EU-US Data Privacy Framework. That page names them and says which is which. If you are in Switzerland: the same clauses apply with the adjustments the Federal Data Protection and Information Commissioner has recognised, and the FDPIC is the authority you can complain to.
How long we keep it
Two different answers, because two different things are held.
Your profile, and what we generated about you. This is ours. As of today: indefinitely. We do not currently run an automatic deletion schedule, so a profile stays in our database until you or we remove it. We are moving to a fixed rule: public profile data re-verified every 24 months and deleted if the channel is gone or the address is dead; AI generated notes and scores deleted or regenerated after 12 months. Until that is running, the deletion route that actually works is the one below. Ask us, and we delete you.
Your reply. This is the studio's. We hold the text of what you wrote for the studio you wrote to, in their campaign, for as long as that studio keeps its account with us, and that can be years. Using up their send allowance or the end of their tracking window does not delete it. What ends it: the studio can ask us to delete any reply at any time, and you can ask us at any time (see "Erase" below); and if nobody from that studio signs in for 12 months, we tell them, and if we hear nothing within 30 days we delete the text of your reply everywhere we hold it (snippet, held message, logs, drafts) and keep only the fact that you replied, and when. If they do respond, the 12 months start again, so a studio that keeps using its account keeps your reply. The software does not yet record sign-ins; the 12 months run from the day it does. When a studio closes its account, it has 30 days to ask for an export of its records, which we deliver within 30 days of the request; the live records are then deleted within 30 days of the later of the end of that window and any deletion request from the studio, nothing covered by an undelivered export is deleted first, and backup copies go within a further 30 days.
What our software made of your reply, and any summary of coverage you published. Also the studio's, kept for as long as their account is open, and deleted with the rest of their campaign when it closes. The sentiment and intent assessment of your reply stays even after the reply text itself has gone under the 12-month rule. A copy of your reply in the studio's own mailbox is theirs and outside all of these rules.
Opt-out records are the exception to everything above. We keep the minimum needed to keep honouring them, your address and handle and the fact, date and scope of what you asked, for as long as we run the platform, because deleting them would let you be contacted again.
Your rights
For anything here, email [email protected]. We answer within 30 days. We do not charge for any of this. To answer an access or deletion request we need to be reasonably sure you are you, so we will reply to the address the request concerns and ask you to confirm from it. If we have any doubt, we may ask for one more thing that only you would have, such as a post from the channel. We will not ask you to send us an identity document.
- Object. Tell us to stop and we stop. For marketing this is absolute. What we stop depends on what you ask for: stop alone ends that campaign to you; naming the game or the studio ends every campaign for that game or from that studio; wanting no contact through Pigeoneer at all ends every studio's and every marketing use of your record.
- Access. Ask what we hold and we will send it to you, including the score, the rationale, the AI notes and any coverage summary.
- Correct. If anything is wrong, including the AI written parts, tell us and we fix it.
- Erase. Ask us to delete you and we delete your database record and start deleting our copies of your reply at once, pass the request to the studio you wrote to within three business days, and finish within 30 days, with backup copies gone at the next rotation and within a further 30 days. We confirm in writing when it is done. If you have also told us to stop, the minimum opt-out record stays.
- Restrict. Ask us to freeze processing while something is disputed.
- Portability. Ask for your data in a machine readable file and we send it, including the scores and notes we generated. We provide that as a service commitment; the statutory right under Article 20 is narrower and depends on conditions that our processing of the pool does not meet.
- Complain. You can complain to a data protection authority without going through us first.
Where to complain:
- UK: the Information Commissioner's Office, https://ico.org.uk/make-a-complaint/, helpline 0303 123 1113.
- EU and EEA: your national supervisory authority. The list is at https://edpb.europa.eu/about-edpb/about-edpb/members_en. In Spain that is the Agencia Española de Protección de Datos, https://www.aepd.es.
- Anywhere else: your local data protection or privacy regulator, if you have one.
If you are in California
The CCPA gives you the right to know what we hold, to delete it, to correct it, and to opt out of any sale or sharing. The categories we hold about you are identifiers, internet or network activity, professional information, and inferences we drew to build a profile of your channel. We collected them from public sources and from your reply if you sent one. We disclose them to the game studio running the campaign, and to the service providers listed at pigeoneer.app/subprocessors. We do not share personal information for cross-context behavioural advertising. Exercise any of these at [email protected]. The "make us stop" section at the top of this page is also a valid opt-out of sale or sharing. We will not discriminate against you for using any of them.
Other pages
- Privacy policy for clients and site visitors: pigeoneer.app/privacy
- Subprocessor list: pigeoneer.app/subprocessors