What Is CBT, Briefly?
Cognitive Behavioral Therapy was developed in the 1960s and 1970s by psychiatrist Aaron T. Beck (depression) and psychologist Albert Ellis (rational-emotive therapy). Its core insight: the way we interpret events — not events themselves — drives much of how we feel and behave. By identifying distorted or unhelpful thoughts and testing them against evidence, people can change emotional and behavioural responses to recurring situations.
CBT is short-term (typically 8–20 sessions), structured, present-focused, and skills-based. It is the most-studied psychotherapy in history. Hundreds of randomised controlled trials and dozens of meta-analyses have established it as a first-line treatment for major depressive disorder, generalised anxiety disorder, panic disorder, social anxiety, OCD (paired with exposure and response prevention), PTSD (with adaptations), insomnia (CBT-I), and several other conditions. The American Psychological Association lists CBT under Division 12's Empirically Supported Treatments for many of these conditions.
Internet-delivered CBT (iCBT) — structured CBT programmes delivered via websites and apps — has its own substantial evidence base. The UK NHS commissions iCBT through its NHS Talking Therapies (formerly IAPT) programme and recommends specific iCBT apps via the NICE Digital Health Technology evaluation framework. Meta-analyses generally find iCBT effect sizes for mild-to-moderate depression and anxiety in the range of d = 0.4–0.7 — smaller than face-to-face CBT but meaningful, especially given accessibility advantages.
What CBT Apps Can and Can't Do
The best AI CBT apps automate structured exercises that map cleanly onto self-help workbooks — thought records, behavioural activation schedules, exposure hierarchies, mood tracking, and basic psychoeducation. They scale well, are available between sessions or when therapy is unaffordable, and can produce real symptom reductions for mild-to-moderate symptoms. They cannot replicate clinical judgement, complex case conceptualisation, transference work, or the therapeutic relationship itself. They are also not designed for crisis, severe symptoms, or trauma processing.
Conditions Where CBT Apps Have the Strongest Evidence
| Condition | Evidence for app-delivered CBT | Best-fit apps in this guide |
| Mild–moderate depression | Strong — multiple RCTs and meta-analyses; iCBT routinely commissioned by NHS | Wysa, Woebot (B2B only), Youper |
| Generalised anxiety disorder | Strong — iCBT consistently outperforms waitlist; smaller-but-real effects vs face-to-face | Wysa, Youper |
| Insomnia (CBT-I) | Strong — CBT-I-specific apps (e.g. Sleepio) have FDA De Novo authorisation; not covered here | Out of scope for this guide |
| Stress / sub-clinical worry | Strong — lower bar; CBT skills generalise well | Wysa, Youper, MindDoc |
| Mild social anxiety | Moderate — CBT works but apps usually lack exposure structure | Wysa with caveats |
| OCD | Weak — OCD requires Exposure & Response Prevention with clinician oversight; general CBT apps are not adequate | None recommended; see licensed CBT therapist |
| PTSD | Weak — trauma-focused CBT or EMDR is the standard; generic CBT apps are insufficient and may be destabilising | None recommended |
| Severe / treatment-resistant depression | Insufficient — needs clinician-led care; apps may complement but not substitute | None as primary intervention |
What Makes a Good CBT App?
Not every app that claims "CBT" delivers genuine cognitive-behavioural therapy techniques. The label is unregulated and "CBT-inspired" prompts sometimes amount to motivational quotes plus mood tracking. A credible CBT app should include:
- Cognitive restructuring: Identifying and challenging distorted thoughts (the core CBT skill)
- Behavioural activation: Structured activities to counter avoidance and withdrawal
- Thought records: Tools for logging automatic thoughts, evidence, and balanced alternatives
- Psychoeducation: Clear explanations of the CBT model and how thoughts influence feelings and behaviour
- Goal-setting and progress tracking: Structured outcome monitoring, ideally using validated instruments (PHQ-9, GAD-7) rather than ad-hoc mood scales
- Clinical validation: Published research demonstrating the app's effectiveness, not just claiming "CBT-based"
See our clinical evidence scorecard for which apps actually have peer-reviewed research and which only claim it. Three out of every four apps marketing themselves as "CBT" have no platform-specific peer-reviewed evidence at all.
How CBT Compares to Other Therapy Modalities
| Modality | Core focus | Best for |
| CBT | Thoughts ↔ feelings ↔ behaviours; cognitive restructuring | Depression, anxiety, insomnia, panic, social anxiety |
| DBT | Distress tolerance, emotion regulation, interpersonal effectiveness, mindfulness | BPD, chronic suicidality, emotion dysregulation |
| ACT | Psychological flexibility; values-based action; defusion from thoughts | Chronic conditions, values conflicts, perfectionism |
| Mindfulness-Based (MBCT, MBSR) | Present-moment awareness; relapse prevention | Recurrent depression, stress reduction |
| EMDR / trauma-focused | Processing traumatic memories with bilateral stimulation or narrative reconsolidation | PTSD, single-incident trauma |
AI CBT apps deliver the cognitive-restructuring + behavioural-activation core of CBT well. They typically deliver a watered-down or surface-level version of DBT, ACT, or mindfulness when they claim those modalities — with one or two exceptions (Wysa's mindfulness pack is more developed than most).
#1 Wysa
Consumer Free CBT, DBT, Solution-Focused Therapy
Clinically validated AI mental health chatbot with 30+ peer-reviewed publications, NHS-endorsed, available in 65+ countries; blends AI self-care with human coaching
- ✓ AI CBT chatbot with 200+ self-care tools
- ✓ Guided meditation, breathing exercises, yoga
- ✓ Text-based coaching with licensed professionals ($19.99/session)
- ✓ Mood tracking and journaling
#2 Youper
Consumer Free CBT, ACT, DBT
Stanford-validated AI emotional health assistant combining multiple evidence-based therapy modalities (CBT, ACT, DBT, PST) with clinically validated outcome assessments
- ✓ AI conversational chatbot (Emotional Health Assistant)
- ✓ Evidence-based CBT, ACT, DBT, PST, Mindfulness interventions
- ✓ Mood tracking and daily check-ins
- ✓ Clinically validated assessments (GAD-7, PHQ-9)
#3 Replika
Consumer Free CBT (informal), Supportive listening
AI companion focused on emotional connection and long-term relationship building rather than structured clinical therapy; persistent memory creates sense of ongoing relationship
- ✓ AI companion with persistent memory and personality development
- ✓ CBT-informed mood coaching and coping skills
- ✓ 3D customizable avatar
- ✓ Voice and video call capabilities (Pro)
#4 Bloom (by Spring Health)
Consumer $14.49 CBT, Mindfulness, Behavioral Activation
Video-first self-therapy approach (700+ interactive sessions) rather than chatbot; acquired by Spring Health — now part of a comprehensive employer mental health platform serving major enterprises
- ✓ 700+ interactive video therapy sessions
- ✓ Self-guided CBT skill-building exercises
- ✓ Journaling and breathing exercises
- ✓ Topics: stress, anxiety, relationships, habits
#5 MindDoc
Consumer Free CBT, Behavioral tracking, Psychoeducation
German-engineered, clinically validated mood tracking app classified as EU Class I medical device; used by 3M+ users globally; focuses on structured psychological assessment over conversational AI
- ✓ AI-powered mood tracking with 3x daily check-ins
- ✓ Clinically validated psychological assessments (PHQ-9, GAD-7)
- ✓ Personalized feedback on symptoms and patterns
- ✓ AI therapist for pattern exploration (added 2025)
Important Limitations of CBT Apps
CBT apps deliver structured self-help exercises, but they cannot replicate the full therapeutic relationship and clinical judgment that a trained CBT therapist provides. Key limitations:
- Cannot perform nuanced case conceptualisation or identify complex cognitive patterns
- Pre-programmed responses may miss important context in a user's situation
- Limited ability to handle therapeutic ruptures, resistance, or complex emotions
- Not validated for severe depression, OCD (which requires specialised CBT protocols like ERP), PTSD, or eating disorders
- Cannot provide the accountability and relational motivation of in-person therapy
Best use: CBT apps work best as a supplement to therapy, a way to practise skills between sessions, or a first step for people with mild symptoms who aren't ready for formal treatment. They are not a replacement for professional CBT with a licensed therapist.
How to Use a CBT App Productively
- Set a specific goal. "Reduce my GAD-7 score by 5 points over 8 weeks" beats "feel better." Apps that include validated outcome instruments make this easier.
- Commit to a cadence. Most iCBT evidence comes from programmes used 2–4 times per week for 6–12 weeks. Sporadic use produces sporadic results.
- Treat it as homework, not therapy. The cognitive-restructuring and behavioural-activation exercises are real CBT skills you are building. They generalise to life off-app over time.
- Pair with a clinician for anything beyond mild symptoms. The strongest evidence base is for app-as-adjunct, not app-as-replacement.
- Re-evaluate at 4 and 8 weeks. If a validated outcome score has not moved by 8 weeks, you likely need a different approach — including possibly seeing a therapist.
When You Should Skip Apps Entirely
- Active suicidal ideation, self-harm urges, or recent attempts — please call 988 (US), Samaritans 116 123 (UK/IE), or local emergency services.
- Symptoms that are getting worse, not better, with self-help
- OCD, PTSD, bipolar disorder, psychosis, eating disorders — all require clinician-led care
- Children and teens — consumer CBT apps have not been validated for paediatric use; the FTC has flagged AI chatbot safety for minors as an enforcement priority
Related Reading