AuraLift vs Woebot: Two Categories, Not Two Brands

Anyone evaluating AI mental health apps in 2026 ends up comparing AuraLift and Woebot eventually. They’re both AI products that hold conversations about how a user is feeling. They’re also, in important ways, different categories of product, with different stated goals, different evidence bases, and different audiences in mind. This is the honest comparison from someone who works on one of them.

Why we’re writing this comparison

Two reasons.

First, the most-asked question we get from prospective AuraLift users with prior exposure to the AI-mental-health space is some version of “how is this different from Woebot?” Woebot is the most recognizable name in the category for English-speaking adults; it’s the reference point most people compare against. We’d rather answer the question directly, including the parts that aren’t flattering to us, than dodge it.

Second, comparison content is too often written as marketing collateral, a thin pretext to make the writer’s product look better than the competitor’s. That isn’t useful to anyone trying to pick a tool. The piece below is built around stated category differences, public evidence, and the actual fit problem from a user’s perspective. Where AuraLift is the better fit, we’ll say so. Where Woebot is the better fit, we’ll say that too.

How each one positions itself

Woebot was founded in 2017 by Dr. Alison Darcy, a clinical psychologist and researcher who came out of Stanford’s digital mental health work. Woebot Health, the company, has consistently positioned its product as a digital therapeutic, closer to a clinical tool than a wellness app, built around the delivery of cognitive behavioral therapy techniques in a chat-based format. The company received FDA Breakthrough Device designation for an investigational version of its product targeting postpartum depression in 2021.¹ The framing has consistently emphasized clinical applicability, partnerships with health systems, and outcomes-driven evaluation.

AuraLift positions itself, by contrast, as an AI wellness coaching product, explicitly not a clinical tool, explicitly not a therapy provider, explicitly not a digital therapeutic. Our public stance, repeated across our app store listings, our website, and our in-product copy, is that LAura is a coach, not a clinician; that AuraLift does not diagnose or treat mental health conditions; and that the appropriate response when a user’s state crosses a threshold is to route to clinical care, not to provide better coaching. We call our target audience the “I’m Fine” generation: high-functioning adults whose needs sit below diagnostic threshold but above the threshold of indifference.

These are different category positions, not different brands inside one category. The clinical-vs-coaching distinction is the spine of the comparison.

Frameworks and method

Both products draw heavily on cognitive behavioral therapy as a method tradition. Both incorporate elements of mindfulness, behavioral techniques, and structured reflection. Below the surface, the methodological emphasis differs.

• Woebot’s public materials emphasize CBT in particular, the surface area of the product is structured around CBT-style conversations, thought records, and activity scheduling, packaged into a fixed therapeutic relationship that the company describes as a “therapeutic alliance with a chatbot.”² The conversational approach is heavily scripted around predefined therapeutic flows.
• AuraLift’s methodological base is broader: CBT plus DBT (distress-tolerance, emotion regulation), ACT (acceptance, defusion, values clarification), and mindfulness, selected for the fit to a coaching, rather than treatment, conversation. The product is built around six emotional registers (Warmth, Reflective, Curious, Calm, Noticing, Empathy) that LAura adapts between based on what’s happening in the session, rather than a single treatment-flow shape.

The practical effect: a Woebot session reads more like a brief, highly structured CBT intervention. An AuraLift session reads more like a conversation with a coach who has access to several frameworks and chooses among them based on what would help. Neither approach is strictly better. They fit different needs.

Audience and use case

Audience is where the divergence is sharpest.

Woebot’s historical use case has emphasized adults with mild-to-moderate depression and anxiety symptoms, including users with diagnosable conditions seeking symptom-focused support. The 2017 randomized controlled trial published in JMIR Mental Health recruited college students with self-reported depression and anxiety symptoms; later studies and partnerships have leaned further into the clinical-population direction (postpartum, substance use, comorbid conditions).³

AuraLift’s audience is explicitly upstream of that. Our target user is high-functioning, often successful, often someone whose answer to “how are you?” is “fine” before any actual answer surfaces. We are not aiming at users in diagnosable distress, and our product surfaces are deliberately built to route those users toward clinical care rather than to keep them inside the coaching loop. The use case is languishing, low-grade chronic stress, the 3am wake-up, the Sunday-night dread, the unprocessed conversation from earlier in the week, not symptomatic depression or anxiety.

These overlap. Some users will fit either product reasonably well. The question is which one is built for the use case you’re actually bringing to it. If you’re looking for a symptom-focused, CBT-driven, possibly clinically-positioned tool, Woebot is closer to your center of gravity. If you’re looking for a daily coaching presence that fits a functional life that wants more depth, AuraLift is closer to ours.

Evidence base for each

Woebot has a longer published evidence base than nearly any consumer-facing AI mental health product. The 2017 RCT in JMIR Mental Health is the most-cited piece, a 2-week study with 70 college students showing significant reduction in self-reported depression symptoms in the Woebot arm versus an information-only control. Subsequent studies have explored postpartum depression, substance use disorder adjuncts, and adolescent anxiety, with mixed but generally positive findings on engagement and symptom outcomes.

AuraLift is newer and has less published evidence at the product level today. We don’t have RCTs in our category yet, and we’d be skeptical of any consumer wellness product that claimed otherwise at our age. What we have, and what we believe matters at our category position, is fidelity to the broader research base on the techniques we draw from (CBT, DBT, ACT, mindfulness, each with extensive published literature), and a Clinical Advisory Board that advises on the product’s safety and clinical-boundary architecture. We treat “effectiveness” as a question we’re honest about not having full answers to yet, rather than a marketing claim. The piece Is AI Coaching Actually Effective goes deeper.

Woebot’s evidence base is the more impressive one as of today. We’re saying that directly because it’s true. What it tells you, though, is that Woebot has more evidence for symptom reduction in people with measurable symptoms, which is consistent with the clinical position the product occupies. It doesn’t directly address the coaching use case AuraLift was built for, where the outcome question is less “did symptoms decrease” and more “did this person move from languishing toward something better.” Different evidence questions for different categories.

Safety architecture

Both products care about safety. The architecture differs.

Woebot publishes a safety approach centered on detection of crisis-related language and routing to professional resources, integrated with the broader CBT-flow shape of the product. Their public materials describe a focus on crisis detection and reasonable clinical safeguards.

AuraLift uses a four-tier risk system: Tier 1 (everyday coaching), Tier 2 (heightened distress, increased monitoring, explicit grounding offers, no clinical claims), Tier 3 (clinical referral, LAura says out loud that this has crossed a line and routes to a clinician), Tier 4 (crisis, hard stop on coaching, route to 988 / 741741 / 911). The system is described in detail in The Four-Tier Risk System Explained.

The conceptual difference: Woebot’s safety architecture is a layer on top of an ongoing therapeutic conversation. AuraLift’s safety architecture is a category-level commitment that the product steps out of coaching mode at risk thresholds. Coaching is explicitly not the right tool past those thresholds, and the product is engineered to admit that.

Voice and feel

Two things that don’t show up in the spec sheet but matter to actual users:

• Voice. Woebot’s voice is warm, structured, slightly playful, and consistent across conversations, a character with an established personality. AuraLift’s voice is adaptive across the six emotional registers, depending on what the conversation calls for; the through-line is what we describe as “a friend who happens to know the frameworks.” Sample sessions of each side by side will read meaningfully differently. Pick the one that fits the way you actually want to be talked to.
• Visual feel. Woebot uses a chat-bubble interface centered on a recognizable robot character. AuraLift centers on a single, calm, glowing orb (no character, no anthropomorphic stand-in), with a coaching surface that’s deliberately quiet and spacious. Different products for different aesthetic temperaments. We notice that users who dislike chatbot-character framing tend to prefer ours; users who find character framing warmer tend to prefer Woebot’s.

When each is the right fit

Honest version, written by the person who works on one of them:

Woebot is the closer fit if:

• You have measurable symptoms of depression or anxiety and you want a structured, CBT-driven tool with the longest published evidence base in the consumer-facing space.
• You like the clinical or quasi-clinical framing, it gives you confidence the tool is serious about outcomes.
• You prefer a chatbot character with a recognizable personality.
• You’re under the care of a clinician and want a between-session symptom-tracking adjunct that speaks the same methodological language.

AuraLift is the closer fit if:

• You’re functional, sometimes very visibly successful, and the issue is that your inner life is flatter or quieter than it should be.
• You don’t want a clinical or quasi-clinical product. You want a coach.
• You’re drawn to a calm, spacious, character-free design rather than a chatbot persona.
• You want a product that’s explicit about not being therapy, and explicit about stepping out of coaching when something has crossed a line.
• You want techniques drawn from a broader base than CBT alone, DBT distress tolerance, ACT defusion, mindfulness, applied flexibly rather than as a fixed flow.

A note on the broader AI mental health landscape

Wysa, Replika, Earkick, Youper, there are a dozen meaningful products in the broader AI mental health space, and each has its own category positioning. The piece you’re reading focused on Woebot specifically because that’s the most-asked comparison from prospective users, but the right way to think about category fit applies across the landscape:

• What is the product’s stated category? (Therapy, coaching, journaling, companion?)
• Who is its target audience? (Diagnosable population, functional adults, kids and teens?)
• What is its method base, and is it disclosed?
• What does it do at a risk threshold?
• What evidence does it claim, and what does the underlying study actually show?
• How does it handle data, training, and retention?

Those questions matter more than any single product comparison. Use them on us, on Woebot, on anyone. The answers should be defensible, in plain language, before any of them earn time with your inner life.

For the broader category context, see What is AI Wellness Coaching and AI Coach vs Therapist. For the deeper effectiveness question, see Is AI Coaching Actually Effective.