We Compared The Features of 126 Mental Health Apps: Here's What We Found

Last updated: May 25, 2026

Mental health apps have a broad self-help baseline, but the features that create clinical depth are still scarce. We built a retained dataset of 126 comparable mental health apps, classified each feature with a seven-label availability scheme, and ran the aggregates to see what buyers can actually access. The goal is to figure out which features matter in mental health apps and what to build if you are shipping your own mental health app.

The dataset spans seven workflow families: self-guided therapy skills, condition-specific treatment, licensed professional care, mood tracking and journaling, crisis and safety support, peer community support, and AI emotional support. For each app we recorded a mental-health-specific feature taxonomy and classified actual access patterns rather than relying on marketing claims alone.

If you want to see what proven feature decisions look like beyond mental health apps, our database of 300 profitable internet businesses breaks down what each one shipped, gated, or skipped.

The comparison table

We built this dataset from scratch. For each of the 126 mental health apps, we inspected public feature information and recorded the availability of 12 feature categories: AI chatbot emotional support, guided CBT and thought reframing, mood and symptom tracking, journaling prompts and diary cards, crisis plans and emergency contact activation, breathing and grounding tools, exposure and ERP exercises, sleep diary and CBT-I coaching, peer communities, licensed therapist or coach access, condition-specific psychoeducation, and clinician dashboard or care-team sharing. Each feature was classified with one of seven standardized availability labels, and the full comparison table is below.

Questions on features of mental health apps

These are the questions we kept returning to while building the dataset. They matter if you are trying to decide which features in mental health apps are table stakes, which ones differentiate, which ones to gate, and what to ship first.

Which features are commoditized in mental health apps?

The commoditized features in mental health apps are psychoeducation, guided CBT, mood tracking, journaling, and breathing tools. Psychoeducation appears in 98.4% of apps, while the other four all sit between 73.0% and 82.5% penetration.

Psychoeducation is the cleanest baseline. It appears in every self-guided therapy skills app, every licensed professional care app, every crisis and safety support app, every peer community support app, and every AI emotional support app in the dataset.

Guided CBT is the next strongest signal because it appears in 104 of 126 apps. In self-guided therapy skills apps, it reaches 97%, which makes it nearly mandatory for products claiming structured self-help.

Mood tracking and journaling form the second part of the category baseline. They are universal in mood tracking and journaling apps by definition, but they also appear broadly across self-guided, condition-specific, licensed-care, and AI emotional support workflows.

Breathing and grounding tools are common enough to feel expected, but not universal enough to define the whole category. They appear in 94% of self-guided therapy skills apps and 91% of crisis and safety apps, which makes them table stakes in those workflows.

The builder rule is simple: a broad mental health app should not launch without educational content, reflective tracking, and at least one structured exercise layer. Missing any of those makes the product feel partial unless the workflow is deliberately narrow.

Which features are usually free by default in mental health apps?

In mental health apps, the features most often exposed for free are psychoeducation, mood tracking, journaling, breathing tools, and crisis planning. Even there, free full access is not dominant: psychoeducation is free full in only 24.2% of present implementations.

Free access is most visible where the feature behaves like public-health support or lightweight self-care. Crisis and safety apps are the clearest example, with crisis planning free full in 73% of present cases and breathing tools free full in 60%.

Consumer self-care features usually use free limited access rather than full free access. Guided CBT, journaling, breathing tools, mood tracking, and psychoeducation all have larger free limited counts than free full counts across the full dataset.

Mood tracking and journaling apps show the dominant freemium mechanic. Mood tracking is present in 100% of that workflow, and 67% of those implementations are free limited rather than fully free.

Self-guided therapy skills apps are the broadest free-access segment, but they still split across free full, free limited, paid, restricted, and unclear labels. MindShift CBT, What’s Up?, MoodTools, and e-couch illustrate the free-full side, while DBT Coach and Sensa Health sit on the paid side.

The pattern for builders is to make the first self-help loop accessible: learning, tracking, journaling, and basic regulation. Full depth, scale, specialist protocols, and human support can sit behind gates without violating category expectations.

Which features are most often limited, paywalled, or premium-only in mental health apps?

The most aggressively gated features in mental health apps are therapist access, clinician dashboards, and AI chatbot support. Therapist access is paid only or restricted in 92% of present implementations, while clinician dashboards are restricted in 71.4% of cases where they exist.

Human support is the cleanest premium layer. Licensed therapist or coach access appears in 50 apps, but only one is free full and one is free limited, which means the category has almost no expectation of free professional care.

Clinician dashboards are gated differently. They are not simply paid in the usual consumer sense; they are mostly restricted by employer, school, provider, institution, care team, or clinical context.

AI chatbot support uses a different gating pattern. It appears in only 20 apps, and half of those are free limited, which suggests caps on conversation depth, usage, or advanced support are the common monetization mechanic.

Specialist protocols are not always paywalled, but they are often hard to access or hard to verify. Exposure and ERP has a 40.0% unclear share among present implementations, while sleep diary and CBT-I coaching splits across free limited, restricted, and unclear access.

Gating in mental health apps therefore operates on three layers: usage limits for consumer self-care, paid access for service-heavy or AI-heavy features, and restricted access for clinical or institutional workflows. A strong product can use all three without confusing buyers if the boundaries are explicit.

If you want to see what premium features look like across 300 different businesses, our database of 300 profitable internet businesses breaks down exactly what each one chose to gate.

Which features are still strong differentiators in mental health apps?

The strongest differentiators in mental health apps are AI chatbot support, clinician dashboards, crisis planning, peer communities, exposure and ERP, and sleep diary or CBT-I coaching. Each one either sits below 40% penetration or is concentrated sharply in one workflow.

AI chatbot support is a differentiator because it appears in only 15.9% of the full dataset. Wysa, Youper, Earkick, Tudle, Sonia, and MindBuddy illustrate the AI-led side, but most mental health apps still do not include conversational emotional support.

Clinician dashboards differentiate institutional and professional products from consumer self-help apps. They appear in 63% of licensed professional care apps and 45% of condition-specific treatment apps, but only 14% of AI emotional support apps.

Crisis planning is a workflow-defining differentiator. It is universal in crisis and safety support apps, yet completely absent from mood tracking and journaling apps in this dataset.

Peer communities set a product boundary rather than a universal quality bar. They appear in 100% of peer community support apps but only 27.0% of the dataset overall, which means their presence signals a specific support model.

Exposure, ERP, fear ladders, sleep diary, and CBT-I coaching are specialist differentiators. They are rare enough that adding them can separate a mental health app from generic self-care, but only when the product has the condition focus to make them credible.

If you're trying to figure out what makes a product genuinely different in its category, our database of 300 proven internet businesses shows how each one carved out its differentiation feature by feature.

Which features are rarely offered in mental health apps?

The rarest features in mental health apps are exposure and ERP at 11.9% penetration, sleep diary and CBT-I coaching at 12.7%, AI chatbot support at 15.9%, and peer communities at 27.0%. These features are not absent because they are unimportant; they are workflow-specific or operationally harder to ship.

Exposure and ERP are the starkest specialist gap. Even inside condition-specific treatment apps, only 26% include exposure, ERP, or fear ladder exercises, which is low for a feature with clear relevance to anxiety and OCD workflows.

Sleep diary and CBT-I coaching has a similar profile. It appears in only 16 apps, with no presence in crisis and safety support, peer community support, and only one AI emotional support app.

AI chatbot support is rare in category terms even though it is highly visible in market conversation. The AI emotional support workflow contains most of the obvious examples, but the feature does not yet define mental health apps as a whole.

Peer communities are also rare outside their own workflow. Mood tracking apps, AI support apps, and crisis tools almost never include them, which suggests community is treated as a core product model rather than a bolt-on feature.

The takeaway for builders is that rare features in mental health apps are not automatically bad MVP bets. The right rare feature can be powerful, but only when it matches the workflow and safety obligations of the product.

Which missing features create the biggest opportunity in mental health apps?

The biggest missing-feature opportunities in mental health apps sit around safety escalation, specialist protocols, and AI-human continuity. Crisis planning appears in only 31.7% of apps, exposure and ERP in 11.9%, and AI chatbot support in 15.9%.

The clearest cross-workflow opportunity is crisis planning. It is universal in crisis and safety support apps, but absent from mood tracking apps and present in only one of seven AI emotional support apps.

That gap matters because mood tracking and AI chat are often where distress signals show up first. A lightweight safety plan or emergency contact workflow could turn those products from reflective tools into safer support loops.

Exposure and ERP create a different opportunity. The feature is clinically meaningful but rare even in condition-specific treatment apps, which leaves room for focused products that make specialist protocols easier to access and understand.

AI-human continuity is the third gap. Mental health apps rarely combine AI chatbot support with licensed therapist access, so the market mostly separates automated support from professional care instead of letting one hand off to the other.

For builders, the best opportunity is not to add every advanced feature. It is to close one high-friction gap that belongs naturally to the workflow: crisis escalation for tracking and AI tools, ERP for condition-specific apps, or care-team handoff for clinical products.

If you want to spot feature gaps that buyers will actually pay to close, our internet business database surfaces the same patterns across 300 different markets.

What should be free versus paid in mental health apps?

In mental health apps, the free layer should cover the basic self-help loop: psychoeducation, mood tracking, journaling, CBT prompts, and breathing tools. The paid or restricted layer should cover human support, clinical oversight, AI depth, specialist protocols, and care-team sharing.

The free layer should let users learn something, record how they feel, reflect on patterns, and complete a basic coping exercise. Those features are common enough that hiding all of them behind payment makes a product feel misaligned with category norms.

Free limited access is the safest default for consumer self-care. Across major self-help features, the free limited share is consistently large, especially for guided CBT, journaling, breathing tools, mood tracking, and psychoeducation.

Free full access makes the most sense for public-health-style tools, crisis tools, and some government or nonprofit products. MindShift CBT, PTSD Coach, Virtual Hope Box, Calm Harm, Safety Plan, and Stay Alive fit that pattern.

Paid access is most defensible where the marginal cost or risk profile is higher. Therapist access, coach access, AI chatbot depth, licensed-care programs, and structured specialist protocols can all be gated without surprising buyers.

Restricted access belongs to features that require context or oversight. Clinician dashboards, care-team sharing, prescription digital therapeutics, school programs, employer programs, and provider-mediated tools should not be treated like ordinary consumer upgrades.

Which features make users upgrade to paid plans in mental health apps?

Users upgrade in mental health apps when they hit limits on self-care depth or need a higher-intensity support layer. The strongest upgrade features are AI chatbot support, therapist or coach access, licensed-care programs, advanced CBT content, and specialist protocols.

AI chatbot support is an obvious upgrade lever because present implementations are usually capped or paid. Youper, MindBuddy, CBT Thought Diary, Clarity, and Deprexis show how conversational or AI-adjacent support can become part of the paid product surface.

Human support is the strongest upgrade trigger by cost structure. BetterHelp, Talkspace, Online-Therapy.com, Brightside Health, Calmerry, and Workit Health all position therapist or professional access as paid rather than free.

Advanced self-guided programs also drive upgrades. DBT Coach, Sensa Health, MoodGYM, Reframe, Inflow, and Ahead App illustrate how structured programs can move beyond free prompts into paid content or subscription access.

Specialist protocols can support upgrades when the condition is specific enough. NOCD, Somryst, Sleepio, and Daylight show how condition-oriented products can justify tighter packaging than general self-care apps.

The upgrade rule is to avoid charging for the first moment of help. Charge when the user asks for more depth, more personalization, more continuity, more accountability, or professional support.

If you're shipping your own product, our database of 300 proven internet businesses includes SaaS examples and the exact features each one chose to gate at upgrade.

What should the MVP of a mental health app include and what should it skip?

The MVP of a mental health app should include psychoeducation, a basic structured exercise layer, mood tracking or journaling, and one workflow-specific anchor. It should skip therapist marketplaces, clinician dashboards, peer communities, AI chatbots, exposure and ERP, and CBT-I unless those features define the target workflow.

The baseline MVP starts with content and reflection. Psychoeducation is present in 98.4% of apps, so an app without educational guidance needs a very clear reason to exist.

The second layer is a concrete self-help action. That can be guided CBT, thought reframing, breathing, grounding, journaling, or symptom tracking, depending on the workflow.

The workflow anchor matters more than adding breadth. A crisis app needs crisis planning; a mood app needs tracking and journaling; an AI emotional support app needs chatbot conversations; a peer app needs moderated groups.

The MVP should skip features that create operational burden before the product has earned it. Therapist access, clinician dashboards, care-team sharing, and peer moderation all add safety, staffing, compliance, or institutional complexity.

It should also skip specialist protocols unless the app is explicitly condition-specific. Exposure, ERP, fear ladders, sleep diary, and CBT-I coaching are too narrow for a general mental health MVP but strong for a focused product.

The best launch shape is a narrow, credible loop rather than a broad mental health super-app. A user should understand what situation the app helps with and what action it helps them take.

If you want to see what an MVP looks like across 300 different businesses that actually shipped and grew, our database of 300 profitable internet businesses lets you copy the patterns directly.

What are other interesting feature patterns in mental health apps?

Beyond the headline patterns, mental health apps reveal several quieter dynamics around packaging clarity, workflow boundaries, and the split between consumer self-care and clinical infrastructure.

Unclear packaging clusters around everyday consumer features rather than only around complex clinical ones. Journaling has 13.8% unclear among present implementations, breathing tools have 14.1%, and mood tracking has 12.0%, which suggests vendors often describe outcomes without stating what users can actually access.

Licensed professional care apps are not always the richest self-guided products. They have 100% therapist or coach access by definition, but exposure and ERP appears in only 5% and sleep diary or CBT-I coaching also appears in only 5%.

Crisis and safety support apps are unusually disciplined. They are strong on crisis plans, psychoeducation, and breathing tools, but they avoid AI chatbots, peer communities, exposure and ERP, and sleep coaching entirely in this dataset.

Peer community apps combine community with professional escalation more often than a simple forum model would suggest. Therapist or coach access appears in 82% of that workflow, but it is paid only or restricted when present.

AI emotional support apps are broader than chatbot wrappers. All seven include guided CBT or thought reframing in some form, which means the AI workflow is usually packaged as a self-help loop rather than a standalone conversation box.