Ruby Chat Review: Privacy Claims vs Play Data Safety
Documentation review · Updated 2026-10-09 · Adults 18+
This Ruby Chat review is built from a due-diligence angle: what the developer states, what Google Play's Data safety panel declares, and where the two do not line up. The method is deliberately narrow. We read the public store listing, the app's own feature copy, the Data safety summary, the dated user reviews, and the developer replies visible on the page as of the evidence capture on 2026-10-09. We did not install the app, pay for a subscription, test voice quality, measure retention, or audit any server. Nothing below should be read as a hands-on performance verdict.
The reason this matters is simple. Ruby Chat markets itself as a private space, and privacy is one of the few claims an AI companion app can make that a buyer cannot easily verify from the outside. The store page gives you two different kinds of statement: marketing language chosen by the developer, and structured declarations filed through Google Play. They are not the same artifact, and treating them as equivalent is the central mistake this review tries to help you avoid.
The buyer thesis: privacy-first due diligence for an AI companion
Most reviews of AI girlfriend apps ask whether the chat feels realistic, whether the voice sounds natural, or whether the character remembers your dog's name. Those are reasonable questions. But for a category that invites intimate disclosure, the more consequential question is what happens to the text, audio, and photos you produce. A companion app can be delightful and still be a poor privacy choice for you. It can also be unremarkable as entertainment and still handle data in a way you find acceptable.
This review therefore treats the product as a data relationship first and an entertainment product second. That does not mean the entertainment side is irrelevant. It means the entertainment side is well covered elsewhere, while the privacy side is usually reduced to a single reassuring sentence. Here, the reassuring sentence is the starting point of the analysis, not the conclusion.
What the developer says about privacy
The listing carries a short privacy promise. It describes conversations as private and never shared, and frames the app as a private space for AI chat and roleplay. That is the developer's characterization. It is worth noting what kind of statement it is: a broad assurance, not a technical specification. It does not name an encryption standard, a retention period, a deletion workflow, a data residency location, or a third-party processor list.
This is not unusual. Store listings are marketing surfaces, and short privacy claims are common across the category. But a buyer doing due diligence should recognize that a sentence like this tells you the developer's intent, not the system's architecture. Intent can be sincere and still be incomplete. The listing also states that data is encrypted in transit and that you can request deletion, which are more concrete, but they are general platform-level statements rather than product-specific guarantees.
What Google Play's Data safety panel declares
The same listing includes a Data safety section. It states that the app may share Location, Personal info, and three other data types with third parties, and that it may collect Location, Personal info, and four other data types. It also states that data is encrypted in transit and that you can request that data be deleted.
Those two statements sit on the same page as the claim that conversations are never shared. They do not necessarily contradict each other, because Data safety categories are broader than chat content. Location and personal information may be collected for analytics, advertising, fraud prevention, or account management without any conversation text being shared. The panel does not tell you which category maps to which purpose, and the summary view does not itemize the underlying flows. That ambiguity is precisely the point.
Why the two statements are not automatically contradictory
A careful reader should resist the temptation to declare a scandal. The Data safety panel uses predefined buckets, and 'Personal info' can include an email address, a device identifier, or a purchase reference rather than the content of a roleplay scene. 'Location' may be coarse and derived from network data rather than precise GPS. The 'three others' and 'four others' are collapsed in the summary and could include diagnostics, app activity, or device identifiers.
It is therefore possible that the developer's promise about conversations is accurate while the app still collects and shares other categories. It is also possible that the promise is narrower than a casual reader assumes, for example covering only certain chat contexts. The documentation does not resolve which reading is correct. A buyer should hold both possibilities open rather than pick the comforting one.
What remains unknown after reading the listing
Several important questions are simply not answered by the public page. Whether chat transcripts are encrypted end-to-end is unknown; the listing only says data is encrypted in transit, which protects data while moving between your device and a server but says nothing about what happens at rest or who can read it. Whether deletion requests remove backups, logs, model-training copies, or derived data is unknown. Whether voice calls are processed in real time by a third-party speech service is unknown. Whether photos generated by the app are stored on the developer's servers or only on your device is unknown.
Also unknown: the retention schedule, the identity of the third parties receiving the declared categories, the jurisdiction where data is stored, and whether any human review of conversations occurs for safety or quality purposes. None of these are accusations. They are gaps. For a privacy-first buyer, gaps are the product decision.
Feature claims and the consumables question
The developer describes voice calls, photo messages, voice messages, many AI characters, memory across sessions, custom personas, roleplay scenarios, eight languages, and premium perks including unlimited voice calls, instant replies, exclusive characters, and no ads. These are developer descriptions, not verified test results. The listing also shows in-app purchases and a Mature 17+ rating with suggestive themes and strong language.
On the premium tier, the word 'unlimited' appears in the developer's copy. Buyers should treat this as a marketing descriptor rather than a contractual measurement. In practice, AI products depend on compute, and compute is a consumable. A subscription that promises unlimited voice calls still depends on servers, rate limits, and fair-use policies that may not be visible on the store page. The honest position is that the listing does not tell you what happens when usage is extreme, and you should ask before assuming.
Dated user reviews: what they can and cannot prove
The page shows individual reviews with dates. A September 2026 reviewer describes the AI losing track of a story after leaving a scene and says the storyline was deleted when backing out. A June 2026 reviewer calls the app fun and worth the money. An August 2024 reviewer reports that tapping AI chat avatars did not work and describes limited use without a subscription. A developer response dated November 2025 says a bug was fixed in the latest version. The page shows an aggregate rating and a review count alongside these.
These are individual testimony at specific moments, not universal proof. A 2024 complaint may describe a version that no longer exists. A 2026 complaint may describe a device-specific issue. Positive reviews may come from users with different expectations. The developer reply is evidence that the team responds, not evidence that every issue is resolved. Reviews are useful for generating questions, not for settling them.
Decision table: how to weigh the evidence
| Signal | What it actually tells you | Weight for a privacy-first buyer |
|---|---|---|
| Developer states conversations are never shared | Stated intent; not a technical specification | Low until backed by a policy or audit |
| Data safety declares sharing of Location, Personal info, and others | Structured disclosure of categories, not itemized flows | High; demands follow-up |
| Data encrypted in transit | Protection during transmission, not at rest or end-to-end | Medium; partial reassurance |
| You can request deletion | A process exists; scope and backups unknown | Medium; verify before relying |
| Dated user complaints about memory and access | Individual experiences at a point in time | Low to medium; question generator |
| Mature 17+ rating and in-app purchases | Audience and monetization context | Contextual; not a privacy signal |
Pros and cons from the documentation
Pros. The listing is transparent enough to include a Data safety panel that declares sharing rather than hiding it. The developer states a privacy intent in plain language. The app supports multiple languages and describes a broad feature set including voice and photos. The developer appears responsive in at least one public reply. The page shows a large install base relative to the review count, which suggests the product is not obscure.
Cons. The privacy promise is broad and unverified. The Data safety panel declares collection and sharing of multiple categories without mapping them to purposes. Encryption is described only as in-transit. Deletion is offered but its scope is unspecified. Some dated reviews describe functional problems with memory and access. Premium perks use absolute language that may not survive contact with real usage limits.
Precise buying questions to ask before subscribing
Ask the developer, in writing, and keep the answer: Does the app process chat content end-to-end, or can the service read messages? Which third parties receive the categories declared in Data safety, and for what purpose? What is the retention period for chat, voice, and photo data? Does a deletion request remove backups and derived data, and how long does it take? Are voice calls handled by the developer or a third-party speech provider? Is any conversation content used to train or improve models, and can you opt out? What are the actual usage limits behind the word unlimited? These are answerable questions, and a privacy-first buyer should not subscribe until at least the first four have clear answers.
Hypothetical examples of how to decide
These examples are illustrative and hypothetical. They contain no invented prices or limits. Suppose you only want light, casual roleplay and you do not share identifying details. In that case, the gap between the marketing claim and the Data safety panel may be tolerable, provided you understand that collected categories like location and personal info may still flow to third parties. Suppose instead you intend to disclose real names, health details, or relationship history in long-running scenes. Then the unknowns about retention, training use, and human review become central, and the store page alone is not enough to proceed. Suppose you are considering a paid tier primarily for voice. Then the consumable nature of voice minutes and the possibility of fair-use limits should be resolved before payment, because the listing's premium language does not answer that question.
How Ruby Chat compares to the general category
Ruby Chat is not unique in making a strong privacy claim while declaring broad data categories. That pattern is common across AI companion apps, and it reflects the tension between marketing and platform disclosure. What is notable here is that the Data safety panel is visible and explicit about sharing, which at least gives a buyer something to interrogate. Many listings are vaguer. The practical takeaway is not that Ruby Chat is uniquely risky or uniquely safe. It is that the category as a whole requires you to read two documents at once: the promise and the declaration.
Frequently asked questions
Does Ruby Chat offer end-to-end encrypted chats? The listing does not say that. It says data is encrypted in transit, which is a different and weaker statement. In-transit encryption protects data while it travels; it does not establish that only you and your device can read the content. End-to-end encryption would need to be stated explicitly, and it is not. Treat this as unknown.
If the developer says conversations are never shared, why does Data safety mention sharing? Because Data safety categories cover more than conversation text. Location, personal info, and other categories may be collected or shared for analytics, advertising, account management, or fraud prevention. The panel does not map categories to purposes, so you cannot conclude that chat content is shared, nor that it is not. The two statements coexist but do not fully explain each other.
Can I delete my data? The listing says you can request deletion. That is a process statement, not a guarantee about scope. It does not specify whether backups, logs, derived data, or model-training copies are removed, nor how long it takes. If deletion matters to you, ask for the specifics in writing before you rely on it.
Are the user reviews reliable? They are individual dated testimony. One reviewer in September 2026 describes memory and story-loss problems; another in June 2026 calls the app fun. A developer reply in November 2025 says a bug was fixed. Reviews help you form questions, but they are not universal proof, and older reviews may describe versions that have changed.
What does the Mature 17+ rating mean for buyers? It signals suggestive themes and strong language, and it indicates the app is aimed at adults. It is not a privacy or safety certification. Buyers should be 18 or older, and the rating should be read as a content descriptor rather than an endorsement of data practices.
What should I do if I cannot get clear answers? Treat the absence of answers as a signal. A privacy-first buyer can simply decline to subscribe and choose a different product whose terms are explicit. You do not need to prove that Ruby Chat is unsafe to decide it is not transparent enough for your use case.
Actionable decision
If your priority is privacy, do not subscribe on the strength of the listing's privacy sentence alone. Read the Data safety panel, note that it declares collection and sharing of Location, Personal info, and other categories, and recognize that the listing does not explain the scope, purpose, or retention of that data. Then send the buying questions above to the developer. If you receive clear, specific answers covering encryption scope, third-party recipients, retention, and deletion, you have a basis to proceed with the paid tier. If you receive vague or no answers, the rational privacy-first decision is to avoid paying and to evaluate alternatives instead, checking each alternative's current terms yourself because this review makes no claims about any other product's features or pricing.