The Top 100 Most Used AI Apps in 2025 - 5th Edition

The Consumer AI Top 100 List

About the List:

• 5th Edition: Published every 6 months since the dawn of consumer GenAI
• Purpose: Track what real consumers are actually using, not just what's hyped
• Global Scope: Analyzes every website and mobile app worldwide

Methodology:

1. Data Sources: Similar Web (websites) and Sensor Tower (mobile apps)
2. Metrics: Monthly visits for web, monthly active users for mobile
3. Selection: Top 50 AI-native companies from each platform
4. Focus: All usage (free + paid), not just revenue

Key Insight:

The list captures consumer imagination, not just commercial success — showing what people genuinely find compelling enough to use regularly.

Timestamp: [0:00-1:08]

New Entrants & Market Evolution

Platform Dynamics:

• Web List: More stable, better for tracking trends
• Mobile List: Less stable due to app store policy changes
• Previous Issues: Early lists dominated by ChatGPT copycats

New Company Statistics:

• 11 new names on web (down from 17 six months ago)
• Ecosystem stabilizing: Fewer dramatic changes
• Mobile cleanup: iOS crackdown removed ChatGPT clones

Emerging Trend - Vibe Coding:

1. Bolt: On brink list (just below top 50)
2. Lovable: Made the web list
3. Replit: Strong showing on web

Category Evolution:

Previously dominated by:

• General chat interfaces
• Companionship apps
• Creative tools

Now expanding into:

• Productivity tools
• Vibe coding platforms
• Developer-focused products

Timestamp: [1:08-3:26]

The Persistent Power of AI Companions

New Companionship Entrants:

1. juicychat.ai - Fresh addition to the list
2. joi.com - New companion platform
3. ourdream.ai - Latest entry in the category

Established Leaders:

• character.ai: Biggest name in companionship
• janitorai.com: Consistent performer
• spicychat.ai: Maintains strong position
• polybuzz.ai: Regular list member
• crushon.ai: Steady presence
• adot.ai: Continuing growth
• candy.ai: Established player

Market Reality:

Despite expansion into productivity and coding, companionship remains a dominant force in consumer AI usage, occupying a significant portion of the top 100.

Timestamp: [3:26-4:06]

Google's Dominant Performance

Four Google Properties Made the List:

1. Gemini (#2 Overall)

• Right behind ChatGPT in rankings
• Web traffic: ~10% of ChatGPT's volume
• Mobile traffic: ~50% of ChatGPT's volume
• User base: Primarily Android users

2. AI Studio (#10)

• Developer-facing sandbox
• Where engineers test Google models
• Surprising consumer-level traffic

3. NotebookLM (#13)

• Narrative violation: Viral September 2024
• Maintained flat or increasing traffic monthly
• Only slight summer academic usage decline

4. Google Labs (#39)

• Consumer-facing sandbox for experiments
• 15% traffic spike when Veo 3 launched
• Multiple experimental products included

Google Labs Products:

• Veo 3: Video generation model
• Doppl: Outfit try-on product
• Whisk: Image generation sandbox
• Portrait: Coaching product
• Project Mariner: Agentic browser sandbox

Timestamp: [4:06-6:23]

Essential Takeaways:

1. Usage vs Revenue: Free usage data reveals true consumer interest better than paid metrics
2. Market Maturation: Ecosystem stabilizing with fewer dramatic changes (11 new vs 17 previously)
3. Category Expansion: Moving beyond chat and creative tools into productivity and coding

Persistent Trends:

• Companionship dominance continues despite new categories emerging
• Google's momentum with 4 properties making the list for first time
• Mobile cleanup as app stores remove low-quality ChatGPT clones

Emerging Opportunities:

• Vibe coding platforms gaining significant traction
• Developer tools achieving consumer-level usage
• Productivity applications finding their market fit

Timestamp: [0:00-6:23]

Primary Source:

• The Top 100 Gen AI Consumer Apps - 5th Edition - a16z's comprehensive report on the most used AI consumer applications

People Mentioned:

• Olivia Moore - a16z partner compiling the Consumer AI Top 100 list
• Justine Moore - a16z partner and co-host discussing AI consumer trends

Companies & Products:

• ChatGPT - Leading AI chat interface maintaining #1 position
• Character.ai - Biggest companionship AI platform
• Gemini - Google's AI achieving #2 position
• NotebookLM - Google's research tool at #13
• Google AI Studio - Developer sandbox in top 10
• Google Labs - Consumer experiments sandbox at #39
• Bolt - Vibe coding platform on brink list
• Lovable - Vibe coding platform on web list
• Replit - Coding platform making the list
• janitorai.com - Established companionship platform
• spicychat.ai - Companionship app maintaining position
• polybuzz.ai - Regular list member in companionship
• crushon.ai - Steady presence in AI companions
• juicychat.ai - New companionship entrant
• joi.com - New companion platform
• ourdream.ai - Latest companionship entry
• adot.ai - Growing companionship app
• candy.ai - Established companion player

Technologies & Tools:

• Similar Web - Data source for website traffic analysis
• Sensor Tower - Data source for mobile app usage metrics
• Veo 3 - Google's video generation model driving Labs
• Doppl - Outfit try-on experimental product
• Whisk - Image generation sandbox tool
• Portrait - Coaching product in Google Labs
• Project Mariner - Agentic browser sandbox

Concepts & Frameworks:

• Vibe Coding - Emerging trend of AI-assisted development platforms
• Consumer AI Native - Companies built specifically for AI-first experiences
• Narrative Violation - When products maintain success beyond viral moments

Timestamp: [0:00-6:23]

Three Categories of Chinese AI Success

Category 1: Built for China, Used in China

These products exist because major Western AI tools like ChatGPT and Claude are banned in China due to regulatory requirements:

• Quark (Alibaba): AI assistant ranking in top 20 on web
• Doubao (ByteDance): Available on both web and mobile
• Kimi (Moonshot AI): General AI assistant in top 20

Why They're Winning:

• Population advantage: China is the world's largest country
• Limited competition: Western products can't enter
• Regulatory moat: Must follow specific Chinese data rules

Category 2: Built in China, Used Globally

Products developed in/around China but can't be used in China:

Leading Examples:

• deepseek: Advanced AI models
• Hailuo AI: Video generation
• KlingAI: Video creation platform
• SEAART.AI: Image generation

Distribution Strategy: Chinese companies distribute through US platforms:

• Krea and Hedra host multiple Chinese models
• Available via subscriptions bundling multiple models
• Access to both closed (Kling, Minimax) and open source (Hunyuan, Qwen)

Category 3: Global Players from China

• Manus: $90 million annualized run rate
• Traffic breakdown: #1 Brazil, #2 USA
• Used both domestically and internationally

Timestamp: [6:25-10:19]

Chinese Dominance in Visual AI

Strength Areas:

1. Single-frame animation: Leading in image-to-video
2. Video generation quality: Competitive with Western models
3. Distribution innovation: Multiple access points

Notable Video Models:

• Seedance: ByteDance model that outscores Veo 3 on some arenas
• Kling: Closed-source premium model
• Minimax: High-quality generation
• Hailuo: Emerging video platform

Distribution Channels:

Consumer-facing:

• Direct websites and apps
• Aggregator platforms (Krea, Hedra)

Developer-facing:

• Fal and Replicate for API access
• One-off runs or integrated APIs
• Both open and closed source options

Strategic Advantage:

Chinese companies excel at both creation and distribution, using US platforms as intermediaries to reach global markets while maintaining domestic presence.

Timestamp: [8:17-9:33]

The Vibe Coding Revolution

Market Performance:

• 6 months ago: Only Bolt on the list
• Now: Bolt on brink list, Lovable and Replit made main list
• Lovable: Recently announced $100 million in ARR

Revenue Retention Analysis:

Using Consumer Edge data to track cohort payment behavior:

Month-over-month retention:

1. Months 0-3: 100% or above revenue retention
2. After month 3: Flattens slightly below 100%
3. Comparison: Extremely rare for consumer/prosumer products

User Behavior Patterns:

• Start with low-cost trial plans
• Actually publish and use their creations
• Buy more credits and upgrade plans
• Suggests either enterprise prototyping or serious solo builders

Why This Matters:

The retention data suggests vibe coding isn't just a trend — users are finding real, sustained value and upgrading their usage over time.

Timestamp: [10:19-12:11]

The Vibe Coding Usage Mystery

Traffic Analysis Reveals:

• Builder traffic > Built product traffic
• More people visiting Lovable/Replit than sites made with them
• Hosted on domains like lovable.app for tracking

Two Possible Explanations:

Theory 1: Professional Usage

• Serious builders buy custom domains
• High-traffic sites no longer show in platform data
• Traffic gets attributed to custom domains instead

Theory 2: Personal Software Revolution

• People building software for themselves
• Not seeking external traffic
• Extremely valuable to individual users
• Shared only with family and friends

Mobile Status:

• No vibe coding apps in top 50 mobile yet
• Several startups emerging but not ranking high
• Companionship apps still dominate mobile
• Next edition may show breakthrough

The Bigger Picture:

Vibe coding might be enabling a new paradigm where individuals create personal software solutions rather than public-facing products.

Timestamp: [12:11-13:38]

Chinese AI Strategies:

1. Three-pronged approach: Domestic-only, global-only, and hybrid products
2. Visual AI dominance: Particularly strong in image and video generation
3. Distribution innovation: Using US platforms as global intermediaries

Vibe Coding Validation:

• Revenue retention above 100% in first 3 months (unprecedented for consumer)
• $100M+ ARR milestones achieved by multiple companies
• Personal software may be bigger use case than public websites

Market Dynamics:

• Brazil emerging as major AI traffic source globally
• Chinese companies successfully exporting AI despite restrictions
• Vibe coding yet to crack mobile but showing strong web growth

Future Indicators:

• Watch for vibe coding mobile breakthrough
• Chinese visual AI models increasingly competitive with Western offerings
• Personal software creation could be the real vibe coding revolution

Timestamp: [6:25-13:38]

Primary Source:

• The Top 100 Gen AI Consumer Apps - 5th Edition - a16z's comprehensive report on the most used AI consumer applications

People Mentioned:

• Olivia Moore - a16z partner analyzing Chinese AI trends and vibe coding data
• Justine Moore - a16z partner with expertise in video AI models

Companies & Products:

• Quark - Alibaba's AI assistant ranking top 20
• Doubao - ByteDance AI assistant on web and mobile
• Kimi - Moonshot AI's general assistant
• DeepSeek - Chinese AI model company
• Hailuo - Chinese video generation model
• Kling - Video creation platform
• SEAART.AI - Image generation from China
• Manus - $90M ARR Chinese AI company
• Krea - Platform hosting Chinese models
• Hedra - Aggregator for various AI models
• Minimax - Closed-source Chinese model
• Hunyuan - Open-source Chinese model
• Qwen - Open-source model gaining popularity
• Seedance - ByteDance video model outperforming Veo 3
• Bolt - Vibe coding platform on brink list
• Lovable - $100M ARR vibe coding platform
• Replit - Coding platform on main list
• Fal - Developer platform for AI models
• Replicate - API platform for AI models

Technologies & Tools:

• Consumer Edge - Data provider for revenue retention analysis
• Veo 3 - Google's video model used as benchmark
• API Distribution - Method for accessing Chinese models globally

Concepts & Frameworks:

• Vibe Coding - AI-assisted rapid application development
• Revenue Retention - Cohort-based payment tracking metric
• Personal Software - Individual-use applications vs public products
• Distribution Strategy - Using intermediary platforms for global reach

Timestamp: [6:25-13:38]

The AI All-Stars: 5 Lists, 2 Years, Consistent Winners

General LLM Assistants:

1. ChatGPT - The undisputed leader
2. Perplexity - Search-focused AI
3. Poe - Multi-model aggregator

Creative Tools Champions:

• Midjourney - Image generation pioneer
• PhotoRoom - Background removal and editing
• Leonardo - AI art platform
• Cutout Pro - Visual editing tools
• Veed - Video editing platform
• ElevenLabs - Voice synthesis leader

Productivity Powerhouses:

• Quillbot - Writing enhancement
• Gamma - Presentation creation

Model Hosting Heroes:

• Hugging Face - Open source model hub
• Civitai - Community model repository

Companionship Constant:

• Character.ai - Present on both web and mobile every edition

The Significance:

These 14 companies have maintained top positions for two full years through massive market volatility, proving their product-market fit and user loyalty.

Timestamp: [13:40-14:30]

The Surprising Truth About Model Ownership

The Initial Assumption:

Two years ago, the prevailing belief was that only companies investing tens of millions in proprietary models would succeed.

The Reality Check:

More than half of the 14 All-Stars don't own their models!

Winning Without Models:

Model Aggregators succeed through:

• Superior UI/UX design
• Deep, specialized workflows
• Community features
• Product experience excellence

Why This Works:

1. Amazing models are now API-accessible
2. Open source options proliferate
3. Product experience matters as much as the model
4. Specialized workflows create differentiation

Key Insight:

In consumer AI, how you package and deliver AI matters more than owning the underlying technology. The wrapper can be more valuable than the model itself.

Timestamp: [14:30-15:08]

Beyond Data: The New AI Network Effects

Traditional AI Network Effect:

More users → More data → Better models → More users

Emerging Network Effects:

Community Model Platforms:

Hugging Face & Civitai:

• Users host models, LoRAs, datasets
• Community comments and rankings
• Exclusive apps only on platform
• First non-data network effects in AI

Voice Libraries:

ElevenLabs:

• Thousands of user-uploaded voices
• More voices → More value for new users
• Community-generated content library

Enterprise Network Effects:

Gamma, PhotoRoom, ElevenLabs:

• Team templates and design libraries
• Organizational knowledge accumulation
• Higher switching costs with investment

The Virtuous Cycle:

Even ChatGPT and Midjourney benefit from user data improving models, which attracts more users, creating a self-reinforcing growth loop.

Why This Matters:

Network effects create defensibility beyond just having the best model — they build moats through community, content, and organizational embedding.

Timestamp: [15:08-17:05]

The Bottom-Up Enterprise Revolution

The New Enterprise Sales Motion:

1. Individual discovers AI tool for personal use
2. Brings to work for professional projects
3. Spreads to teammates organically
4. Multiple users adopt individual plans
5. Company consolidates to enterprise contract

Success Stories:

• ElevenLabs: Individual users become internal champions
• Gamma: Teams adopt templating systems
• PhotoRoom: Design libraries spread through organizations

Why It's Working:

Traditional Enterprise Sales:

• Top-down approval from 10+ stakeholders
• Convince employees to adopt
• Long, complex sales cycles

AI Bottom-Up Approach:

• Self-serve enterprise sales
• Users already convinced of value
• Natural organic spread within teams
• 10-20 individual users trigger enterprise conversation

The Prosumer Bridge:

AI enthusiasts aren't just tourists — they have real use cases and become powerful internal advocates for adoption.

Viral Loop:

Team adoption → Personal use at home → Share with friends/family → More consumer growth → More enterprise opportunities

Timestamp: [17:05-18:29]

From Chaos to Stability: The AI Market Matures

The Early Days (First 2 Lists):

• 50% turnover between editions
• Total chaos in rankings
• Traffic fluctuating wildly month-to-month
• Founders angry about 2-week data lags
• Rankings changing faster than data could capture

What Worked Then:

Only three categories showed staying power:

1. Companionship apps
2. General LLM assistants (ChatGPT-style)
3. Creative tools

The Random Spikes:

• Products would appear one month
• Spike in traffic dramatically
• Disappear the next month
• No sustainable growth patterns

Now (5th Edition):

• Market stabilizing significantly
• Only 11 new entries (vs 17 last time)
• Consistent performers maintaining position
• Clear category leaders emerging
• Sustainable business models developing

The Evolution:

From experimental playground to mature market with established players, proven use cases, and predictable growth patterns.

Timestamp: [18:29-19:38]

The Durability Test:

1. 14 companies maintained top 100 status across all 5 editions
2. 2+ years of consistent performance despite market volatility
3. Model ownership not required - over half use others' models

Network Effects Emerging:

• Community platforms (Hugging Face, Civitai) creating model ecosystems
• Voice libraries (ElevenLabs) building content moats
• Enterprise embedding through templates and team adoption

Enterprise Transformation:

• Bottom-up adoption replacing traditional top-down sales
• Individual users becoming enterprise champions
• Self-serve model leading to organic team spread

Market Evolution:

• From 50% list turnover to increasing stability
• Three early categories (companionship, LLMs, creative) remain dominant
• New categories (productivity, vibe coding) gaining traction

Success Factors:

• Product experience matters as much as model quality
• Deep workflows and UI excellence create differentiation
• Community features and network effects build defensibility

Timestamp: [13:40-19:38]

Primary Source:

• The Top 100 Gen AI Consumer Apps - 5th Edition - a16z's comprehensive report on the most used AI consumer applications

People Mentioned:

• Olivia Moore - a16z partner analyzing AI All-Stars and market evolution
• Justine Moore - a16z partner discussing network effects and enterprise adoption

Companies & Products:

• ChatGPT - Consistent #1 AI assistant
• Perplexity - Search-focused AI All-Star
• Poe - Multi-model aggregator platform
• Character.ai - Companionship leader on web and mobile
• Midjourney - Image generation All-Star
• PhotoRoom - Background removal and editing tool
• Leonardo - AI art platform
• Cutout Pro - Visual editing All-Star
• Veed - Video editing platform
• ElevenLabs - Voice synthesis with voice library
• Quillbot - Writing enhancement tool
• Gamma - Presentation creation platform
• Hugging Face - Open source model hub
• Civitai - Community model repository

Technologies & Tools:

• LoRAs - Low-Rank Adaptations for model customization
• Voice Library - ElevenLabs' community voice collection
• API Models - Accessible AI models via APIs
• Open Source Models - Freely available AI models

Concepts & Frameworks:

• AI All-Stars - Companies present in all 5 top 100 editions
• Non-Data Network Effects - Community and content-based moats
• Bottom-Up Enterprise Sales - Consumer-led enterprise adoption
• Prosumer Growth - Individual users driving team adoption
• Model Aggregation - Success without proprietary models

Timestamp: [13:40-19:38]

The Five-Edition Journey

Edition 1-2: The Wild West

• 50% list turnover between editions
• Traffic fluctuating wildly month-to-month
• Only 3 working categories: companionship, LLM assistants, creative tools
• Random products spiking and disappearing

Edition 3: Expansion Begins

• First appearance of video generation
• Music creation tools emerge
• Moving beyond just text and images
• Market starting to find new use cases

Edition 4-5: Market Maturation

• 11 new entries (down from 17)
• Clear All-Stars established
• Newcomers clustering around themes
• Stabilization of core categories

The Pattern:

New categories emerge in clusters — just like vibe coding appeared across multiple companies simultaneously, suggesting market timing matters more than first-mover advantage.

Key Insight:

The AI consumer market has evolved from experimental chaos to structured growth with predictable patterns and established winners.

Timestamp: [19:40-20:36]

The Surprising Grok Success Story

The Debut:

• #4 on web list in first appearance
• Strong showing on mobile as well
• Appeared near end of measurement period

Growth Drivers:

1. Companion features attracting users
2. Image generation capabilities
3. Video features drawing engagement
4. Social platform integration potential

Market Reality:

Even with ChatGPT's dominance, the market has room for multiple winners with different strengths.

Other Late Entrants:

• Meta AI making a dent on web (not mobile yet)
• Shows the market isn't locked up
• Different use cases supporting different platforms

The Takeaway:

Late entry doesn't mean lost opportunity — specialized features and unique positioning can still capture significant market share quickly.

Timestamp: [21:08-21:52]

Future Predictions for Consumer AI

Productivity/Prosumer Tools:

Where hallucinations are a problem, not a feature:

• Spreadsheet generators
• Financial modeling tools
• Slide deck creators
• Email automation

Why they'll explode:

• Model reliability finally sufficient
• Workflow/UI reaching maturity point
• Revenue already exploding

Expected New Entrants:

• Perplexity Comet (as separate domain)
• GenSpark - Search alternative
• Manus - Already showing strong metrics

Category Expansion:

As foundation models improve (GPT-5, Claude 4, etc.):

• Better at math and logic
• Reduced hallucinations
• Naturally supercharges accuracy-dependent products

The Tipping Point:

Products moving from "cool but unreliable" to "good enough to depend on" will drive massive adoption in professional use cases.

Timestamp: [21:52-23:40]

The Untapped Opportunities

Education (EdTech):

• Requires accuracy (low hallucination tolerance)
• Huge market potential
• Model improvements enabling viability

Personal Finance:

• Absent from both web and mobile lists
• Expected explosion on mobile particularly
• Trust and accuracy finally achievable

AI-Native Social Platforms:

• Facebook becoming "boomer AI images"
• Room for multiple AI-first social networks
• Grok potentially pioneering this space

Health & Wellness:

• Consumer health questions
• Structured coaching programs
• Therapy and mental wellness
• Currently handled by ChatGPT but deserves specialized products

The GPT-5 Effect:

Model improvements in health understanding could trigger a wave of specialized consumer health products.

Key Insight:

Many use cases currently served by general assistants will spin out into specialized vertical products as models improve.

Timestamp: [23:40-24:50]

The Beautiful Randomness of Consumer AI

The Unpredictability Factor:

• Vibe coding: Nobody predicted it 9 months ago
• Now multiple companies with $100M+ ARR
• Shows how quickly new categories emerge

The Pattern:

If breakout categories were predictable:

• Hundreds would be building them
• Competition would be instant
• Success would be harder

Historical Surprises:

1. Companionship apps' massive scale
2. Vibe coding's sudden emergence
3. Chinese AI's global reach
4. Bottom-up enterprise adoption

Future Certainty:

There will be 1-2 completely unexpected categories in the next few editions that nobody is talking about today.

The Lesson:

Consumer AI rewards innovation over prediction — the biggest opportunities are in categories that don't exist yet.

Timestamp: [24:50-25:14]

Market Evolution:

1. From chaos to structure: 50% turnover → 22% turnover
2. Category expansion: Text/image → video/music/coding
3. Verticalization: General tools → specialized solutions

Emerging Trends:

• Productivity tools reaching reliability threshold
• Model improvements enabling new categories
• Vertical unbundling from general assistants

Untapped Categories:

• EdTech - accuracy finally viable
• Personal Finance - trust barrier lowering
• AI-Native Social - new paradigms emerging
• Health & Wellness - specialized coaching/therapy

Surprises & Constants:

• Grok's #4 debut shows market still open
• Companionship remains dominant force
• Unexpected categories will emerge (like vibe coding did)

The Big Picture:

Consumer AI is transitioning from experimental playground to mature ecosystem with specialized tools for every use case.

Timestamp: [19:40-25:37]

Primary Source:

• The Top 100 Gen AI Consumer Apps - 5th Edition - a16z's comprehensive report on the most used AI consumer applications

People Mentioned:

• Olivia Moore - a16z partner compiling the Consumer AI Top 100
• Justine Moore - a16z partner providing analysis and predictions
• Anish Acharya - a16z partner who wrote about AI product verticalization

Companies & Products:

• Grok - Debuted at #4 on web list
• Meta AI - Starting to gain traction on web
• Gemini - Used for different purposes than ChatGPT
• Claude - Specialized use cases
• ChatGPT - Different strengths than competitors
• Perplexity - Search-focused AI
• Viggle - Strong revenue metrics
• Perplexity Comet - Potential future list entrant
• GenSpark - Expected search alternative
• Facebook - Mentioned as "boomer AI images" platform

Technologies & Tools:

• GPT-5 - Next OpenAI model improving accuracy
• Claude 4 - Anthropic's model series
• Foundation Models - Base models powering applications

Concepts & Frameworks:

• Verticalization - AI products specializing for specific use cases
• Hallucination Problem - Accuracy issues in productivity tools
• AI-Native Social - Social platforms built for AI-first experiences
• Bottom-Up Enterprise - Consumer-led business adoption
• Category Emergence - Unpredictable new use cases appearing

Timestamp: [19:40-25:37]