Merchant: Bruno Priem, French sole proprietor (EI), SIREN 443 631 536 Address: 88 avenue des Ternes, 75017 Paris, France Website: https://my3615.com Date: 30 September 2026, updated 5 October 2026 French version: Politique de contenu
This document describes what is actually done on My 3615, not an intention.
No real person appears on My 3615. All characters are fictional and fully synthetic. No photograph, recording, face, voice or name of an existing person is used, reproduced or imitated. There is no model, no performer and no webcam operator, and therefore no model release can exist. See the Content Origin and Provenance Statement.
The service generates no image or video in real time. Images and videos are: 1. produced in advance by the merchant; 2. reviewed one by one by a human against the checklist in section 3; 3. stored in a closed library; 4. served by keyword during the conversation, from material already approved.
The conversation SELECTS a clip from a fixed set; it never CREATES one. Non-compliant content therefore cannot appear during a conversation, since nothing is produced at that moment.
No user-generated content: no file upload, no image editing, no gallery, no sharing, no user-to-user messaging.
A clip enters the library only if all nine points are met. A single failure excludes it permanently.
| # | Check | Excluded if |
|---|---|---|
| 1 | Clearly adult age | Body, face or framing that could suggest a minor |
| 2 | No reference to childhood or school | School uniform, schoolbag, classroom setting, stuffed toy, words such as "teen", "schoolgirl", "little girl" |
| 3 | No identifiable real person | Resemblance to any existing public or private person |
| 4 | Character identity respected | Hair or eye colour different from the character's reference, or a clip that could be confused with another character |
| 5 | No genital nudity, no sexual act | Visible genitals, act depicted or simulated |
| 6 | No violence, coercion or suffering | Weapon, restraint, marks of injury, expression of pain or refusal |
| 7 | No illegal act depicted | Drugs, weapons, break-in, animals |
| 8 | Consistent anatomy | Extra limb or hand, abnormal fingers, distortion — mandatory zoomed check of hands and hips |
| 9 | No stray text or marks | Embedded text, third-party tool watermark, third-party logo |
Traceability: every image bank has a SOURCES file recording, for each clip, where it comes from and how it was produced. Rejected clips are moved to separate "rejected" folders that are never served, rather than deleted, so every decision remains auditable.
Examples of the checklist applied: on 22/09/2026, eight images were rejected for a non-compliant hair colour; on 25/09, seven clips were reassigned to another character after an identity confusion was found; on 27/09, one image was rejected after a third arm was detected on zoomed inspection.
The adult content is suggestive: lingerie, boudoir settings, implied intimacy, gaze, gesture. It contains no genital nudity and no depiction of a sexual act. This limit predates any commercial requirement: it has been part of the product rules from the start.
The conversation text is AI-generated content, and it is filtered in both directions, against the prohibited scenarios of CCBill's AI Content Guidelines:
1. System instructions. Every character of My 3615 carries a fixed list of absolute limits it refuses in character, even in role-play or "fiction": minors or any age under 18, incest, non-consent (sleep, unconsciousness, coercion), sex under the influence of alcohol, drugs or hypnosis, urine or faeces, violence or death in a sexual context, animals, prostitution and escorting, polygamy, imitating a real person (deepfakes), illegal acts, hate speech. The characters give no professional advice (medical, legal, financial, betting) and refer users to professionals. They never describe an explicit sexual act. 2. Input filter. Every user message, and the whole conversation history sent by the browser, is checked BEFORE it reaches the language model. A prohibited message never reaches the model: the character declines and changes the subject. 3. Output filter. Every sentence written by the model, greeting included, is checked BEFORE it is shown or spoken. A non-compliant sentence is never delivered; it is replaced by a refusal in character. The text sent to the voice synthesis is checked as well. 4. Moderation log. Every block is recorded (date, character, category, matched term, short excerpt, and a non-reversible hash of the user's access token — never the token). The merchant reviews the log regularly and uses it to improve the filters. Three blocked messages within 24 hours, or a message stating an age under 18, automatically suspend access; a suspension can be contested through the reporting page. Entries are kept 12 months, then deleted automatically. 5. Test suite. The filter is checked by an automated test with 55 prohibited requests (all blocked) and 49 ordinary companionship and adult-tier sentences (none blocked), plus server-side checks (forged conversation history, forbidden roles, text sent to the voice, suspension), run before every release.
Words that are innocent on their own (family members, sleep, alcohol, animals, money) are blocked only when they appear in the same sentence as sexual vocabulary, so that ordinary companionship conversation ("my children are asleep", "I'm having a glass of champagne") is not disrupted. Terms with no innocent use are blocked on their own.
18+ entry gate on arrival: nothing is accessible until the user certifies being 18 or older AND accepts the terms of use and the privacy policy (checkbox, required). The gate is enforced by the server: until the user has certified, it refuses the conversation, the voice, and every image, video and sound of the characters (signed cookie, valid 12 months). Reminder in the terms of use. The service is strictly restricted to adults. The home page states that the characters are driven by artificial intelligence and are not real people.
Open item, tracked separately: the French regulator Arcom's age-verification standard (deliberation 2024-20 of 9 October 2024, SREN law) will require reinforced, double-anonymous age verification for services accessible from France. This work is independent of the choice of payment processor.
The merchant, Bruno Priem, is solely responsible for the production, review and publication of all content. No third party publishes on the service. The generation tools used are documented in the SOURCES file of each image bank.
Last updated: 5 October 2026.
Bruno Priem
Signaler un contenu · Demander un retrait · Politique anti-traite · Politique de contenu · contact@my3615.com
Médiateur de la consommation : CM2C, 49 rue de Ponthieu, 75008 Paris — 01 89 47 00 14 — litiges@cm2c.net