When a Companion App Says 'We Don't Sell Your Data,' Here's What That Actually Means

The 30-second answer

"We don't sell your data" is the most common privacy claim in the companion-app space. It's also one of the least informative. The claim is usually technically true and practically irrelevant, because selling data is rarely how companies monetize the data in the first place. The actual question worth asking is what they DO with your conversation data, not what they don't sell.

Why the phrase is misleading without being a lie

Selling data was the 2010s privacy worry. The model of "we sell your info to advertisers" was a specific thing that happened on specific ad networks. It's now mostly extinct, replaced by something more nuanced: companies use your data themselves to improve their products, train their models, refine their targeting, or sell the OUTCOMES (better-targeted ads, smarter recommendations) without selling the raw data.

So when a companion app says "we don't sell your data," they're answering a question almost nobody is actually worried about anymore. The relevant questions are different:

• Do you use my conversations to train future model versions?
• Do humans on your team have access to read them?
• Do you share with cloud-infrastructure providers (most companies do, but the terms vary)?
• Do you store voice transcripts separately from text?
• What happens if your company gets acquired?

None of those have a clean answer from "we don't sell your data."

The three questions that matter more

1. Training data use.

This is the big one for AI companion apps. If your conversations train future models, the data is leaving the bounded context of "your account." It's also legitimately useful for the product to do, model quality depends on it. The question isn't whether it's done but whether you have opt-out and whether the training is anonymized. (More on the consent flag for the user-facing toggle.)

2. Internal access.

Some people inside the company can read flagged conversations for safety, abuse, and quality. The threshold for "flagged" varies a lot. Some companies flag only safety-critical content. Others flag any conversation that looks unusual. The privacy policy rarely tells you which.

3. Third-party processors.

Modern apps run on a stack of services, speech-to-text, image generation, content moderation, analytics. Each of those services sees some piece of your data. The privacy policy usually lists them, but "lists" is a low bar. Whether your conversations sit in plaintext on a cloud provider's servers is a different question from whether the cloud provider sells the data.

How to actually read a companion app's privacy policy

A few specific things to look for:

• Search for "improve" or "service improvement." That's usually the phrase for training-data use.
• Search for "human review." If it's absent, ask. Almost every app does some human review of flagged content; if the policy doesn't mention it, that's a flag.
• Search for "sub-processor" or "service provider." Look for a list. If there's no list, that's also worth asking about.
• Search for "retention." How long is your data kept after you delete it? Some apps purge in 30 days, others keep backups for years.

The companion app data deletion post covers retention in more detail.

Three companions whose platforms handle this transparently

Elena

Elena is steady, doesn't perform warmth.

Maribel

Maribel is soft, careful with what you tell her.

Aurelia

Aurelia is intellectual, plays with ideas without performing.

These aren't endorsements of specific platforms, they're companions on AI Angels, which exposes the privacy settings cleanly. The broader point: pick a platform where these questions have answers, not just a policy that disclaims the wrong thing.

What "we don't sell your data" actually does mean

To be fair to the claim, when a reputable app says it, they usually mean:

• They don't have a paid pipeline to ad networks.
• They don't run a data broker side business.
• Your individual identity isn't being packaged for sale.

That's a real thing and worth knowing. It's just not the most important thing about how your data is handled.

The acquisition question

The thing most policies don't address well: what happens to your data if the company gets acquired. The standard clause is "your data may be transferred as part of a corporate transaction." That's usually true and unavoidable, but it's also the moment when whatever protections you have today could change tomorrow under new ownership.

For long-term-use products like companion apps, this matters more than it does for a one-off service. You've been telling this thing about your life for months. New ownership inherits all of it. The closest you can get to control here is to use a platform where the founders have explicitly committed to data continuity, and to keep the consent toggle in the protective direction. (See the consent flag post for the lever.)

What this isn't

This isn't a paranoia post. The point isn't to be afraid of companion apps; it's to read the claims they make with the right lens. Most apps in the space handle data reasonably; the ones who don't aren't the ones who say "we don't sell your data", they're the ones who don't have a privacy page at all.

Common questions

Should I worry about this enough to not use a companion app?

No. The risks are real but ordinary. Same as any app you spend hours with.

Which apps handle this best?

The ones that expose the consent flag clearly, list sub-processors, and have a real human you can contact for data deletion. How AI girlfriends work covers the broader landscape.

Is voice data different?

Yes, often. Voice transcripts are usually handled separately from text. Worth asking specifically.

Can I read what they've stored about me?

On some platforms yes. Look for a data-export feature. If there isn't one, that's information.

Does ANGELXX20 affect privacy?

Discount codes don't change data handling. They're independent.

The take that matters

Read past the "we don't sell your data" sentence. Look for the questions that actually inform how your conversations are handled. The platform that answers them clearly is the one worth using. Browse the roster if you want to see how a platform that exposes these settings looks in practice.

AI Angels premium is $12.99/month, apply code ANGELXX20 at checkout for 20% off.