On paper, OurDream AI covers most of the basics: a named US company, third-party age verification where the law requires it, a written ban on real-person likeness, self-service account deletion and an unusually detailed Trust & Safety site. The weak spots are that your chats can be used to train its AI models, your messages pass through outside AI providers, and public-character images go into a shared pack. Here is what the official pages said on Oct 27, 2026.
This is a test drive. I read OurDream’s terms, privacy policy, refund and content policies, help center and Trust & Safety site. I did not pay for a plan, and nothing below is a claim about hands-on use. For prices and features, see the OurDream AI test drive, and for how it ranks against the rest, best AI girlfriend apps.

The short answer
OurDream is a real company with public, dated policies, and it explains how moderation works and who can see your chats. That is more than many adult AI sites publish. I found no breach for ourdream.ai listed on Have I Been Pwned when I checked on Oct 27, 2026.
But “safe” has two parts. The platform side looks organised. The data side asks for trust: the privacy policy allows training on your content, outside model providers process your messages, and automated systems scan private chats. If you write as though a company system reads every message, you are using it the way its own pages describe.
Who runs it
The terms of service (last modified August 26, 2026) name Dream Studio USA, Inc., a Delaware corporation at 1111B S Governors Ave, Dover. The refund, underage and content policies also name Dream Studio LLC and Tektopia Ltd as part of “Dream Studio”. Where Tektopia Ltd is registered is not stated.
The terms cover ourdream.ai and ourdream.io as one service with one account. They also include an arbitration clause and class action waiver. You can opt out within 30 days of first accepting the terms by writing to [email protected] or the Dover address.
Age verification
Everyone confirms they are 18 or older at sign-up, per the underage policy. On top of that, the Trust & Safety age verification page (last modified June 29, 2026) says identity verification is required where local law requires it, naming:
- The UK, under the Online Safety Act
- Australia, under the eSafety Commissioner’s expectations
- Any other jurisdiction whose law requires it
The check is run by Go.cam or AgeVerif. Methods can include AI age estimation, a live selfie, a government ID or credit-card verification. OurDream says it receives only a pass or fail signal and never receives your ID, selfie or video. It stores the “verified 18+” signal on the account and session details for up to 30 days in a cookie.
That is a sensible design, and naming the providers is better than most sites manage. Outside those regions, the pages describe only the sign-up age confirmation.
Payments
The help center lists cards and crypto, and G2A codes if a card is declined. PayPal is not accepted. The privacy summary says your card is held by the payment processor, not OurDream, though the processor is not named.
Two things to know before paying:
- Crypto payments are final and non-refundable (refund policy section 4).
- The name on your card statement is not published.
Refunds are tight: the help center says nothing is refunded once more than 25% of a product is used. The full rules, and how to stop billing, are in how to cancel OurDream AI.
Your chat data
This is the part to read slowly.
Training
The privacy policy (section 2, last modified August 26, 2026) lists using information to improve the services, “including to train our artificial intelligence/machine learning models”. The information it collects includes your chats. I found no opt-out for training on the pages I checked. It also says OurDream may use de-identified or aggregated data for any purpose.
Who can see your chats
The Trust & Safety privacy summary answers this directly:
| Who | Sees your private chats? |
|---|---|
| You | Yes |
| Other users, including character creators | No |
| OurDream staff | Only with your explicit support consent, or for a lawful legal request |
| Third-party AI providers | Yes, to generate the reply, under their data-processing agreements |
The moderation page adds that automated classifiers process all content, including private chats and private generations. Human reviewers see content only if it is flagged, reported or part of an investigation. OurDream says reviewers “do not browse private chats”.
The outside providers are the part you cannot check. They are not named, so you cannot read their policies.
Images
Private generations stay private, per the same summary. But images you generate with a public character are added automatically to that character’s public pack, visible to all users, with the information about who generated them removed. If that bothers you, use private or unlisted characters.
Uploading photos of people is not allowed at all. The help center says reference photos are blocked “due to possible deepfakes”.
Security and sale
The privacy policy says data is encrypted at rest and in transit and hosted on a SOC 2 Type 2 certified platform. It also says no transmission is fully secure. The Trust & Safety site says “We do not sell your chats”, and the privacy policy says OurDream does not currently sell personal information as Nevada law defines it. The site uses Google Analytics, and the privacy policy says it does not respond to Do Not Track signals.
Deleting your account
Deletion is self-service. The help center gives the path: your profile page, the “danger zone” at the bottom, then delete my account. The privacy policy names Settings → Account management instead.
Per the help center, deletion takes 24 to 48 hours and is permanent. It removes generations, unpublished characters, chat history and unused dreamcoins, and cancels an active subscription.
Two limits. A public character that becomes popular may be kept after you delete your account (privacy policy section 9). And information can be kept for legal reasons. For other data requests, email [email protected]. The privacy policy aims to answer within one month.
You can also delete single chat messages from the three-dot menu, which the safety tools page says removes them from the conversation and the model’s context.
Content rules, real people and deepfakes

The prohibited content policy (last modified June 9, 2026) bans content designed to resemble, depict or impersonate real people, including public figures and celebrities, without consent. The Trust & Safety site goes further: no celebrities, no private individuals and no fictional characters clearly modelled on a real person. The terms ban voice impersonation.
The same policy bans content involving or resembling minors, non-consensual scenarios, incest and a list of extreme categories. A written real-person ban like this is the minimum I need before covering a service here.
Moderation runs in layers: input and output classifiers, behaviour signals, human reviewers and user reports. Appeals for content go through the Help Desk to a different reviewer, and account suspensions to [email protected].
What users report
I have no third-party user complaints to report. Trustpilot has removed the ourdream.ai profile, and Reddit could not be reached by the research tools.
One complaint comes from OurDream itself. Its help center says the content-flag system is “a bit too sensitive right now” and may flag safe content, and that it is kept strict to satisfy payment providers. Expect some false flags.
What is not published
To be fair to OurDream, these are gaps in what I could verify, not proof of a problem:
- Which AI providers process your chats
- A way to opt out of model training
- The payment processor and the billing descriptor
- Where Tektopia Ltd is registered
- Identity checks outside the UK, Australia and other places where the law requires them
My verdict, on paper
OurDream publishes more about safety than most adult companion sites. Named age-check providers, a clear table of who sees your chats, self-service deletion and a firm real-person ban are all real positives.
The trade-off is your words. Training is allowed, unnamed outside providers process every message, and classifiers read everything. Treat chats as not private from the company or its vendors, and keep identifying details out.
Safer habits if you use it
- Keep real names, places and identifying details out of chats and characters.
- Use private or unlisted characters if you do not want your generations in a public pack.
- Sign up with an email you use only for this kind of site.
- Pay by card if you might need a refund, since crypto is final.
- Cancel billing, then delete the account when you are done.
If you want alternatives

- Secrets AI says chat history is deleted immediately when you delete your account, and its card charges show as “S LABS INC”. Its privacy policy also allows using data to train its AI models.
- Candy.ai names its statement descriptor (“EverAI”), but account deletion is by email request rather than a button.
- Lovescape bans celebrity look-alikes. Its privacy policy, last modified June 4, 2024, does not explicitly mention training on chats, and it describes no ID-based age check.
All three were checked against their official pages on Oct 25, 2026. For a side-by-side, see Candy.ai vs OurDream AI.
Read next
- Is Nectar AI legit?: a scam check for another companion site
- Is Kindroid safe?: encryption, age checks and deletion
| # | Service | Score | From | |
|---|---|---|---|---|
| #1 | 9.0 | €3.99/mo | Visit site | |
| #2 | 8.8 | $9.17/mo | Visit site | |
| #3 | 8.0 | $9.99/mo | Visit site |
FAQ
Is OurDream AI safe to use?
On paper, mostly. It is run by Dream Studio USA, Inc. in Delaware, uses Go.cam or AgeVerif for age checks where the law requires them, bans real-person likeness and offers self-service deletion. The main concerns are training on your content and unnamed outside AI providers processing your chats.
Does OurDream AI read your chats?
Automated classifiers process all chats. Staff see private chats only with your support consent or for a lawful legal request, and human reviewers see flagged or reported content. Third-party AI providers process your messages to generate replies.
Does OurDream AI need ID?
Only where the law requires it, including the UK and Australia. There, Go.cam or AgeVerif checks you by face estimation, selfie, ID or credit card, and OurDream receives only a pass or fail.
Does OurDream AI train on my chats?
Its privacy policy says information, including chats, may be used to train its AI models. I found no opt-out on the official pages.
How do I delete my OurDream AI account?
Go to your profile page, scroll to the "danger zone" and click delete my account. The help center says it takes 24 to 48 hours, is permanent and cancels any active subscription.
