Version 2 · 14 October 2026 · for creators
A short, honest inventory. Not a cookie-banner document, but an actual list of what is in the database with your name on it, why it’s there, and how to get rid of it. Here to listen rather than create? See what we store about listeners.
Your account: your handle, a bcrypt hash of your password (never the password itself), your account status, when you signed up, and when you were last seen. An email field exists on the record and is currently optional.
Your profile: bio, avatar image, social links, and status line. All of it is optional, all of it is public once you’ve published something.
Your work: the audio files, the script you submitted, the title and description. Stored in a private bucket; the public only ever gets time-limited links, and only to works you have published.
Moderation records: for every upload: how closely a machine transcript of the audio matched your script, the automated policy screen’s verdict (including any short phrase it flagged), the human decision, the reason, and a hash of the script at each step. Append-only: these are never edited or deleted, because their whole purpose is to be a record that can be trusted. They do not contain your audio, your script or the transcript.
The transcript itself: the machine transcript of your audio, kept with your work so the reviewer can read it. It is deleted with your work, not kept with the moderation records.
Your studio work: drafts, chat history and projects you keep in Script Studio and Score Studio, so they are there when you come back.
Keys you connect: if you add your own Lovense developer token or AI key, it is stored so the tools can use it on your behalf. It is never shown back to you in full, only its last four characters.
Activity events: sign-ins, uploads, publishes, and tool usage, each with your creator id and a timestamp. This is the metering substrate for the eventual billing. It records that you used a tool and when, not what you made with it.
Your acceptance of the creator agreement: which version, when, a hash of the exact text you were shown, and a one-way hash of the IP address it came from. The IP is hashed, not stored, so it can confirm an acceptance was real without being a location log.
Reports: if someone reports your work or profile, the report is stored with its reason and outcome. Reports are filed anonymously; the reporter is stored as a one-way hash, so even the operator cannot see who filed one.
Your audio and scripts are read and heard by the operator, as part of review. That is the one unavoidable human access, and section 4 of the agreement says so plainly.
Four service providers touch your data because the platform runs on them, and a fifth only if you connect it:
| Who | What they get | Why |
|---|---|---|
Supabase (EU, eu-west-1) |
The database and the private audio bucket | It is the database |
| Railway (EU West) | The running application | Hosting |
| OpenAI | Your audio file, transiently, for transcription | Moderation ground truth |
| OpenRouter | Your script and transcript text, transiently, and your Script Studio conversations | The automated policy screen, and Script Studio’s writing assistant |
| Lovense, only if you connect a token | Device commands sent under your Lovense developer account, when you preview haptics or someone pairs a toy to your work | Driving toys through Lovense’s app |
Nothing is sold, and nothing is shared with anyone else. No advertisers, no data brokers, no other creators.
In the EU: Supabase in eu-west-1 and Railway in EU West. The transcription,
screening and writing-assistant providers process in the US, transiently, on the audio
and text you send them. Lovense is a separate company with its own privacy policy.
Your account, works and transcripts: until you delete them. Moderation and acceptance records: kept after deletion, for the reasons in section 7 of the agreement. They hold decisions, reasons, timestamps and hashes, not your audio, script or transcript. Activity events: kept as the billing substrate.
A written retention schedule with actual numbers on it is an open task, not a finished one. When it exists it will be linked here.
Ask the operator directly, and you’ll get:
There’s no ticket system. It’s a person you know, and the answer comes back from her.
This describes what the software actually does today, checked against the code rather than aspirationally. It has not been through a GDPR review. That’s on the list, alongside the solicitor. If you spot something here that worries you, raise it; this document is meant to make that possible, not to close the subject.