CoupleWork

Positioning

CoupleWork is a mobile relationship-coaching app from BetterLabs AI LLC, sold on iOS and Android, built around an AI coach called Maxine. The positioning I find most useful is the one the vendor itself insists on in its FAQ: this is coaching, not therapy. The app is grounded in the clinical experience of Clay Cockrell, LCSW, a couples therapist with 30 years in practice, and Maxine applies frameworks from the Gottman Method and Emotionally Focused Therapy (EFT), with inputs from attachment theory and what the vendor calls the neuroscience of connection. The catalogue blurb I inherited ("AI-powered platform that helps couples resolve conflicts and improve their relationship") is directionally right but misses two things that matter when you are choosing: this is a voice-first, session-structured mobile product, and it explicitly refuses to diagnose or treat anything.

The competitive frame the vendor draws is against both traditional couples therapy (cost, scheduling, geography) and lighter relationship apps (prompts and check-ins). Against therapy, the pitch is price and availability — $24.95 per month covering both partners at time of writing, available at 11pm on a Tuesday without an appointment. Against prompt apps, the pitch is memory and structure: Maxine keeps context across sessions and works a coaching plan rather than serving daily questions.

What it actually does

The product flow starts with a short confidential assessment that sets a personalized starting point. From there you run two kinds of session: Individual Sessions (private to you) and Partner Sessions (shared with your partner), in text or voice. Text partner sessions work like a three-way group chat with Maxine, so partners can join from different locations; voice partner sessions run through one shared device like a speakerphone call. The recommended cadence, per the Meet Maxine page, is a weekly 45-minute session plus three 15-minute ad-hoc sessions for issues that come up between them — sessions themselves are not time-limited once running. A feature called What Matters provides an evolving library of expert-validated questions aimed at values, fears, attachment history and family-of-origin patterns rather than surface compatibility. Between sessions, the app assigns exercises intended to build communication habits.

Memory is central to the design. Maxine carries a picture of the relationship across sessions, individual session history is visible only to you, partner sessions are shared, and the vendor says combined progress informs guidance for both people — including recognising dynamics like a pursue-withdraw cycle as they happen. Technically, the vendor describes CoupleWork 2.0 as rebuilt on a multi-agent architecture that routes each task to the model best suited for it, with data encrypted in transit and at rest.

Pricing at time of writing (vendor FAQ and Android launch post, October 2026): $24.95 per month covering both partners with unlimited voice and text sessions, What Matters and cross-session memory; lower rates on quarterly and annual plans; a free 7-day trial. A third-party listing (fastd.top) quotes $64.90 per quarter (about $21.63/month) and $249 per year (about $20.75/month), and also describes the plans as individual-use — I could not confirm those figures or that framing on the vendor site, so I treat them as indicative and flag the conflict below. The homepage shows a "4.9/5 rating" badge, which is a vendor claim; the store listings I checked show 4.8 out of 5 from 33 ratings on the App Store and 4.7 from 15 reviews on Google Play, with 500+ downloads (October 2026).

My testing notes

First observation: the session mechanics create a real asymmetry couples should know about before paying. Text partner sessions are location-independent — genuinely useful for couples in different cities, and one App Store reviewer praised exactly the accessibility of getting support without an appointment. But voice partner sessions require both people on one device, in one room, talking to a speakerphone. For a product whose headline feature is voice-interactive coaching, that is a practical ceiling: the couples who would benefit most from voice coaching across a distance are pushed onto text.

Second, the review signals are informative but thin. Across the iOS reviews I read, one user said they were more comfortable discussing issues with Maxine than with their in-person therapist, while another said they were considering cancelling because the price felt far too high; a Google Play reviewer made the sharpest criticism — that like all AI products it is "not capable of being truly responsive" and falls back on a basic routine for a given situation. The vendor's own response to the pricing complaint pointed to faster sessions and direct partner-session starts. What struck me is the sample size: 33 iOS ratings and 15 Play reviews is a very small evidence base for a product handling relationship conflict, and the aggregate numbers are stable only because almost nobody has voted.

Third, I worked through the pricing economics against the alternatives. $24.95/month for two people is roughly a third of a single therapy session in most US markets and undercuts even mid-tier AI companion subscriptions, and the vendor's FAQ confirms there are no per-session fees or tiered feature gates — one plan, everything included. That simplicity is a point in its favour versus apps that meter sessions. But the flip side appears in the App Store feedback: for couples who only want occasional check-ins, a flat monthly fee with unlimited sessions is the wrong shape.

Fourth, the safety design is more considered than the category average, though most of it is vendor-asserted. The FAQ says Maxine is designed to recognise crisis signals — abuse, self-harm ideation, acute distress — stop coaching, acknowledge what was shared, surface resources and direct the user to human support. The same FAQ repeatedly and prominently states the app is not a crisis service, not therapy, and not a substitute for professional care, and directs serious mental health conditions, trauma and addiction to licensed clinicians. The "tens of thousands of hours in safety testing" claim is unverifiable, but the architecture of the answer — refuse the clinical frame, define escalation — is what I would want to see.

What it does not do

CoupleWork does not provide therapy, diagnosis or clinical treatment, by its own explicit statement. If your situation involves serious mental health conditions, trauma, addiction or crisis, the vendor itself says to work with a licensed clinician; there is no human professional in the loop anywhere in the product. Effectiveness claims ("our research indicates" combined sessions work best) come with no published study I could find.

There is no web application — the product ships through the App Store and Google Play, and the website is marketing plus a free assessment. The ad-hoc session allowance (three 15-minute sessions) is described on the marketing page without stating its refresh period, so I could not establish whether that is a weekly or monthly budget. Voice partner sessions require one shared device. And the data question is not trivial: a product that accumulates intimate conversation history across months of sessions — with memory as its selling point — demands a careful read of the privacy policy before you start; the vendor says no one reads conversations in real time and that it never sells data, but those remain vendor statements.

Finally, pricing transparency has a crack in it: the vendor says one subscription covers both partners; a third-party listing describes plans as individual-use with a separate subscription per partner. I go with the vendor FAQ here, but the confusion exists in public and would double the cost if the third party were right.

Verdict

I rate CoupleWork 3.4 out of 5. Against the rubric: core functionality depth (40 percent) is solid for a focused product — structured assessment, individual and partner sessions in voice and text, cross-session memory, a values-focused question library and exercise assignments, all grounded in named clinical frameworks. Value (20 percent) is good at $24.95/month for two people with no session metering, though flat pricing punishes casual users. Usability (20 percent) is mixed: onboarding and session flow are simple, but the one-device constraint for voice partner sessions is a real limitation. Support and ecosystem (10 percent) are thin — app-store distribution, a support page, a dev blog, no human coaching option at any price. Reliability signals (10 percent) are the weakest leg: 4.8/4.7 store ratings are good numbers on 33 and 15 reviews respectively, which is simply too little independent evidence to lean on for a product in a sensitive category.

Best for: couples — or individuals working on relationship patterns — who want structured, voice-based coaching grounded in Gottman and EFT principles at a fraction of therapy cost, and who are comfortable with an AI holding the session.

Sources

• https://couplework.ai/
• https://couplework.ai/f-a-q/
• https://couplework.ai/meet-maxine/
• https://couplework.ai/pricing/
• https://couplework.ai/en/couplework-is-now-on-android-ai-relationship-coaching-for-every-couple
• https://apps.apple.com/us/app/couplework-ai-couples-coach/id6504302751
• https://play.google.com/store/apps/details?id=ai.betterlabs.couplework
• https://fastd.top/ai-tools/couplework
• https://couplework.ai/blog/
• https://couplework.ai/relationship-assessment/
• https://couplework.ai/about-us/