Important: Woebot Is No Longer Available to Consumers
Woebot Health permanently shut down its consumer-facing app on June 30, 2025. If you are looking for an AI mental health companion you can download today, this review is informational only — you cannot sign up for Woebot directly. We have left this review up because Woebot's history is the most important case study in the AI mental health space, and because the reasons it failed shape how we think about every other platform on this site. For active alternatives, see our Wysa review, Headspace review, or our best mental health apps guide.
Overview: What Was Woebot?
Woebot was an AI conversational agent built around cognitive behavioral therapy (CBT), interpersonal therapy (IPT), and dialectical behavior therapy (DBT) frameworks. It was founded in 2017 by Dr. Alison Darcy, a Stanford clinical psychologist, with the explicit thesis that an AI agent could deliver structured therapeutic techniques in a way that was both scalable and clinically rigorous. Unlike general-purpose chatbots, Woebot's responses were not generated by a large language model in the modern sense — they were grounded in scripted clinical protocols and triggered by a constrained NLP system designed to be safe by construction.
At its peak, Woebot was the most clinically validated digital mental health tool in existence, with 14 randomized controlled trials and peer-reviewed publications in JMIR and other journals. It was used by hundreds of thousands of consumers and integrated into employer wellness programs and health system pilots. Then, in June 2025, the consumer app was shut down. The company pivoted entirely to enterprise and payer licensing, and the most-studied AI therapy product in the world disappeared from app stores.
Why Woebot Mattered Clinically
Woebot's clinical evidence base was unmatched. Most AI mental health tools cite testimonials and Trustpilot scores as social proof. Woebot cited 14 RCTs. Studies showed measurable reductions in symptoms of depression and anxiety after as little as two weeks of daily use, comparable in some endpoints to early-stage human-delivered CBT. The methodology was published, the protocols were grounded in established therapy modalities, and the founder had real clinical credentials.
For users with mild-to-moderate symptoms who could not access or afford traditional therapy, Woebot was a meaningful intervention — not a substitute for a therapist, but a structured daily practice that taught real CBT skills. Its scripted, multi-choice response format, which felt game-like and limiting to some users, was the same property that made it safe: the AI could not say something dangerous or unvetted because it could not say anything that wasn't pre-written by clinicians.
Why It Shut Down
The official explanation cited evolving FDA regulatory expectations for AI mental health tools and the difficulty of maintaining a consumer product under that regulatory burden. The reality is more layered.
First, the regulatory environment did genuinely tighten. As large language models entered the mental health space in 2023-2024, the FDA signaled growing scrutiny of AI tools making implicit or explicit therapeutic claims. For a small clinical-grade product, the cost of compliance with a moving regulatory target was substantial.
Second, Woebot's scripted approach felt increasingly dated next to ChatGPT-era conversational AI. Users who tried Woebot after using GPT-4 reported it felt limiting and game-like. The constraint that made Woebot safe also made it feel clinically conservative in a market where competitors were promising free-form empathetic conversation.
Third, the unit economics of consumer mental health are brutal. Woebot was free at the point of use, with revenue coming from B2B partnerships. As the cost of cloud inference and clinical content updates rose, the consumer app became a cost center that the enterprise business could no longer subsidize.
Key Features (Historical)
- AI CBT chatbot with conversational therapy sessions
- Daily mood check-ins and mood tracking graphs
- Exercises grounded in CBT, IPT, and DBT frameworks
- Crisis detection and safety monitoring
- 24/7 availability with empathetic NLP responses
- Randomized controlled trial-validated methodology
What Woebot Got Right
Woebot pioneered the idea that an AI mental health tool should be evaluated by the same standards as any other clinical intervention. It published its evidence. It used established therapy frameworks rather than inventing new ones. It built a crisis detection layer that escalated to human resources. It treated safety as a design constraint, not a marketing claim. Every credible AI mental health product today inherits these standards, even when it does not credit Woebot for them.
What Woebot Got Wrong
Weaknesses
- ✗ Consumer app permanently shut down as of June 30, 2025 — no longer B2C accessible
- ✗ Scripted, multi-choice response format felt game-like rather than genuinely conversational
- ✗ False safety flags disrupted conversation flow for non-crisis users
- ✗ Inflexible recommendations that couldn't accommodate personal circumstances
- ✗ Frustration from long-term users locked out after access code changes
- ✗ FDA regulatory burden cited as key reason for B2C exit
The Lesson
A clinically rigorous AI mental health product cannot survive on rigor alone. It also has to be conversationally pleasant, sustainably priced, and able to absorb regulatory cost. Woebot was the best in the world at the first thing and underweighted the rest.
What to Use Instead
The closest active replacement is Wysa, which combines a CBT-grounded AI chatbot with an optional human coaching tier and holds an FDA Breakthrough Device Designation for chronic pain-associated depression. Wysa's free tier gives you most of the Woebot experience without paywalling the fundamentals, and its premium tier at $99.99 per year is meaningfully cheaper than therapy alternatives.
For users who want AI plus a structured meditation library, Headspace's Ebb companion is the most clinically thoughtful generative AI in the consumer mental health space, although it is primarily deployed via enterprise channels. For users who need actual licensed human therapists rather than an AI companion, BetterHelp remains the largest network. For crisis support, call 988 immediately — no AI tool is appropriate for crisis intervention.
Verdict
Woebot was the most important AI mental health product of its era and the clearest cautionary tale for the products that came after it. Its shutdown is not a verdict on the category — it is a verdict on what happens when clinical rigor is not paired with conversational warmth, sustainable economics, and a regulatory strategy. We continue to recommend reading Woebot's published research as the baseline for evaluating any AI mental health claim.
See our Wysa vs Woebot comparison for a deeper look at how the active leader compares to the legacy benchmark.