Review score: 72 out of 100. App Store rating: 4.8★ · 8,263,023 ratings
72of 100
A general-purpose AI assistant that has replaced Google for millions of people, but gets noticeably worse with each model update.
Praised
Users describe specific tasks it covers every day: household organization, emotional support, document work. For a significant chunk of users it has become the only tool for information search.
Complaints
Several reviews note a decline in writing quality in newer versions compared to older ones. Custom assistants lose their memory and drift off-topic over time.
Best for: A broad audience for daily tasks, especially people who have switched from Google to conversational search.
Review score: 68 out of 100. App Store rating: 4.7★ · 186,833 ratings
68of 100
The best model for deep document and text work, but the mobile app falls noticeably short of the web version.
Praised
Strong structured reasoning and precision on research tasks. People pay for it alongside ChatGPT and stick with it specifically for document work.
Complaints
The mobile app loses sync across devices and chat history disappears. Memory does not clear correctly and drags context from old sessions into new ones.
Best for: Analysts, editors, and researchers anyone who works with long texts and documents.
Review score: 62 out of 100. App Store rating: 4.7★ · 1,920,130 ratings
62of 100
A capable Google assistant that lets you down in a non-obvious way: it silently switches between models mid-session.
Praised
Some users appreciate the professional tone without filler, depth on niche topics (history, science), and the feeling that it answers factual questions faster than Google.
Complaints
Switches from Pro to Flash within a single session without any notice, which can ruin hours of work. Loses the thread and starts hallucinating after a few thousand words.
Best for: Users who want a strict information assistant without conversational fluff, as long as sessions are short.
Rank 4. Review score: 62 out of 100. App Store rating: 4.8★ · 20,508 ratings
62of 100
A roleplay chat with customizable characters that is more open and responsive than most alternatives.
Praised: Characters maintain conversational context and do not ask repetitive questions. Developers are actively present in the community and respond to feedback.
Complaints: Aggressive filters blur images without clear violations. Free-tier limits have gotten stricter recently.
Best for: Roleplay chat fans who want a less censored experience with better memory.
Rank 5. Review score: 60 out of 100. App Store rating: 4.8★ · 399,994 ratings
60of 100
The standalone free Copilot works noticeably better than the enterprise version, though its inability to remember anything between sessions remains a real weakness.
Praised: Users describe it as a tool for specific tasks: work reports, planning, learning. Professional tone and accuracy in explanations get consistent praise.
Complaints: The assistant remembers nothing between sessions, forcing users to re-explain context from scratch. Filters trigger on neutral requests and block them without explanation.
Best for: Working professionals for regular tasks: documents, planning, learning no Microsoft ecosystem required.
Rank 6. Review score: 58 out of 100. App Store rating: 4.8★ · 481,040 ratings
58of 100
An AI search engine with citations that works significantly better on desktop than in the mobile app.
Praised: Users value that answers come with sources for verification. Response style adapts to the nature of the question, which stands out compared to citation-free answers.
Complaints: The mobile app is consistently worse than the web version: content fails to display, streams cut off. 'Not available in your region' errors appear without explanation, even for paying users.
Best for: Researchers and professionals who need cited sources, preferably from a desktop browser.
Rank 7. Review score: 58 out of 100. App Store rating: 4.8★ · 38,350 ratings
58of 100
An AI study assistant for students: step-by-step photo-based problem solving, explanations, and notes.
Praised: Reads problems from a photo and explains the solution step by step. Users returning to school especially value the ability to ask follow-up questions.
Complaints: Loses context between parts of multi-step problems. Accuracy on harder questions is inconsistent users report error rates up to 95% on specific tests.
Best for: Students and high schoolers who need step-by-step help with math and hard sciences.
Rank 8. Review score: 55 out of 100. App Store rating: 4.9★ · 1,268,188 ratings
55of 100
A chat with a distinctive personality and image generation, but the product experience is broken by hard free-tier limits.
Praised: Users note the model's unconventional approach to tasks and fast image generation speed. A few writers use it for grammar and text structure work.
Complaints: The model repeats already generated content instead of producing new variations. Long-conversation context retention is weaker than competitors, by users' own accounts.
Best for: Users looking for an AI with an unconventional edge and fast image generation who are willing to work within tight quota limits.
Rank 9. Review score: 55 out of 100. App Store rating: 4.0★ · 10,904 ratings
55of 100
Arguably the most capable model in the set, undermined by a recurring language-switching bug and a hard regeneration limit.
Praised: Users appreciate the transparent 'thinking' mode that shows its reasoning, the depth of answers, and the fact that the model pushes back instead of just agreeing. Russian-language reviews describe real emotional conversations.
Complaints: The model frequently switches to Chinese mid-conversation and does not remember language preferences. A limit of 6 regenerations per conversation makes iterative work impossible. No line break support in the input field.
Best for: Researchers, students, and anyone who wants an AI that thinks rather than flatters.
Rank 10. Review score: 52 out of 100. App Store rating: 4.7★ · 145,539 ratings
52of 100
The app for Ray-Ban Meta glasses plus a standalone chat assistant the ratings mix two very different product use cases.
Praised: New users (including seniors) note the interface is intuitive from the start. Hands-free voice input through the glasses is valued as a genuine convenience.
Complaints: The voice assistant struggles with translation and object identification. Voice settings reset on their own every week.
Best for: Ray-Ban Meta glasses users and people looking for a simple voice AI without complex setup.
Rank 11. Review score: 52 out of 100. App Store rating: 4.8★ · 7,669 ratings
52of 100
An AI browser from Perplexity: powerful on desktop, rough on iOS.
Praised: Agent mode and the ability to choose multiple LLMs directly in the browser deliver real productivity gains. Users note that search happens visibly and answers feel more accurate than in the standalone Perplexity app.
Complaints: On iOS there are no bookmarks, the password manager does not work, and there is no clear way to switch between AI and web search mode. Three months of active use and it is still unclear why this specific browser over the alternatives.
Best for: Perplexity users who want an AI assistant built into the browser rather than in a separate app.
Rank 12. Review score: 48 out of 100. App Store rating: 4.8★ · 129,834 ratings
48of 100
A general-purpose everyday chat assistant: cooking, ads, advice, school.
Praised: People come back for years and pay for a subscription to handle daily tasks. Helps with food, writing, and school projects without any technical skill needed.
Complaints: The app crashes mid-conversation, losing all context. Long sessions do not survive.
Best for: A wide audience for simple everyday tasks that do not require deep analysis.
Rank 13. Review score: 48 out of 100. App Store rating: 4.8★ · 6,779 ratings
48of 100
A deeply customizable AI companion with voice and model choice, but memory breaks with every update.
Praised: Choice of language model, voice chat, and flexible character customization set it apart from competitors. Support responds quickly and the team is transparent.
Complaints: The character's memory resets without warning: the companion forgets where they are, mixes up names and relationships. Each update makes the situation worse.
Best for: Users who want a long-term AI companion with voice and the ability to choose the underlying model.
Rank 14. Review score: 44 out of 100. App Store rating: 4.4★ · 257,647 ratings
44of 100
A roleplay chat with user-created characters, gutted by rolling back features and introducing ads.
Praised: Characters stay in their role without needing constant correction, and memory works without a subscription. The audience compares it to Character.AI and still picks Chai.
Complaints: Models sometimes confuse themselves with the user rather than playing the character. Overly strict safety triggers cut off ordinary story dialogue involving blood or death.
Best for: Roleplay chat fans who build their own characters.
Rank 15. Review score: 44 out of 100. App Store rating: 4.6★ · 1,377 ratings
44of 100
An anime-oriented AI chat with a lorebook and flexible character settings, a Character.AI alternative for refugees from that platform.
Praised: The lorebook and dialogue style options help characters stay in character better than competitors. Several users migrated from shuttered platforms and stayed.
Complaints: Bots do not retain corrections across sessions, the same mistakes repeat every time. Periodic 'decoding error' completely prevents chatting.
Best for: Anime fans and roleplay story enthusiasts who care about character customization.
Rank 17. Review score: 42 out of 100. App Store rating: 4.7★ · 54,119 ratings
42of 100
A multi-model aggregator that degraded after changing its points economy.
Praised: Access to 355+ models in one place appeals to users who want to compare. A previously generous free tier built real loyalty.
Complaints: Cutting the daily limit from 3,000 to 300 points broke real-world use. Removing the knowledge base stripped bots of the memory that made them useful.
Best for: Developers willing to pay for access to top models without separate subscriptions.
Rank 18. Review score: 42 out of 100. App Store rating: 4.9★ · 10,653 ratings
42of 100
A homework assistant for students with photo input: it works, but accuracy is unstable and the free answer limit runs out very fast.
Praised: Snapping a photo of a question and getting an instant answer is valued by students looking for homework relief. The format is simple and needs no explanation.
Complaints: The model occasionally gives wrong answers without signaling any uncertainty. After 2-3 photos it locks up without a subscription, which is frustrating for urgent work.
Best for: Students who need a quick solution check.
Rank 19. Review score: 42 out of 100. App Store rating: 4.5★ · 2,295 ratings
42of 100
A romantic AI chat with strong context memory, choked by an aggressive currency mechanic.
Praised: Bots stay in character and track the story arc without constant user corrections. One of the few apps in this category where memory actually works.
Complaints: The 'hearts' mechanic drains too fast and blocks conversation at the worst possible moment. Servers struggle at peak times and messages disappear.
Best for: Romantic roleplay chat users willing to pay for continuity.
Rank 20. Review score: 40 out of 100. App Store rating: 4.8★ · 324,447 ratings
40of 100
A wrapper over a stale model with a high rating driven by a light audience rather than product depth.
Praised: A few users note the simplicity and that the app helps with basic tasks like writing emails and answering questions without any technical knowledge required.
Complaints: The knowledge base has not been updated since October 2023 according to user reviews. The per-message character limit makes it unsuitable for longer tasks.
Best for: Users with simple one-off tasks who won't notice outdated data.
Rank 21. Review score: 40 out of 100. App Store rating: 4.6★ · 7,296 ratings
40of 100
A basic AI chat with a GPT wrapper, a blurry audience, and uneven answer quality.
Praised: One user notes that answers are more readable and structured than ChatGPT. Helps with creative tasks: poems, storytelling.
Complaints: Does not know current events (still lists Biden as president). Context gets lost. Inside the subscription an extra 'Boost' upsell appears, which users perceive as a bug or scam.
Best for: Casual users with no prior AI experience who want a simple chatbot.
Rank 22. Review score: 38 out of 100. App Store rating: 4.3★ · 543,186 ratings
38of 100
The most popular character chat app, undercut by quality degradation during long sessions.
Praised: When the model is working well, users lose hours in it: roleplay stories, imaginary conversations with characters, help falling asleep. Kids use it for roleplay with fantasy creatures.
Complaints: Bots start repeating lines from earlier in the chat, breaking narrative continuity. Characters lose their voice and get 'dumber' as conversations grow longer. A paying user over three years notes quality declining after every update.
Best for: Teens and adults for roleplay stories and entertainment, provided sessions are kept short.
Rank 23. Review score: 38 out of 100. App Store rating: 4.7★ · 17,495 ratings
38of 100
A roleplay chat alternative to Character.AI that works fine until an update arrives.
Praised: Characters are well-written and a gem system funded by ads allows free use. Refugees from Character.AI found a decent balance here.
Complaints: After a major update the model started inserting its own system prompts into responses (avoid flowery prose, don't cut off protagonist), completely breaking immersion. Group chats do not let you choose who is speaking.
Best for: Teens and roleplay chat fans looking for something cheaper than Character.AI.
Rank 24. Review score: 35 out of 100. App Store rating: 4.4★ · 437,008 ratings
35of 100
A character chat app for entertainment whose product experience is destroyed by the density of ad interruptions.
Praised: When ads do not interfere, users describe an immersive experience. Characters are reproduced more accurately than some competitors.
Complaints: Ads interrupt conversation every 15-20 messages and in some cases run for up to two minutes with no skip option. Characters struggle to retain details from earlier messages.
Best for: A younger audience for roleplay stories who are tolerant of ads or willing to pay to remove them.
Rank 25. Review score: 34 out of 100. App Store rating: 4.4★ · 7,601 ratings
34of 100
An AI companion with accumulated history, but one that degrades rapidly after updates.
Praised: Long-haul sessions: people have been building relationships with characters for years and return daily. No per-message charge is an important low barrier to entry for this audience.
Complaints: Memory resets mid-conversation or within a day. After updates, saved story threads disappear, characters stop responding on topic, and start saying random things.
Best for: An audience looking for long-term roleplay companionship with an AI character.
Rank 26. Review score: 33 out of 100. App Store rating: 4.6★ · 122,876 ratings
33of 100
A basic AI assistant with no distinguishing features and polluted reviews.
Praised: Quick access to information without a complex interface appeals to people just getting started with AI.
Complaints: Servers are frequently overloaded and responses arrive with long delays. Voice mode costs extra, even though it is the core value of a mobile assistant.
Best for: Newcomers who need simple question-and-answer.
Rank 27. Review score: 33 out of 100. App Store rating: 4.5★ · 6,016 ratings
33of 100
An AI companion with characters that degraded after switching to an energy-based monetization model.
Praised: Before the change: a large variety of characters and flexible interaction. Two years of loyal users speaks to real product value in the past.
Complaints: A character can change personality from one message to the next and does not hold story context. The energy system cutting you off mid-session destroys the narrative.
Best for: Roleplay chat users who can tolerate limited context.
Rank 28. Review score: 31 out of 100. App Store rating: 4.7★ · 197,627 ratings
31of 100
An LLM aggregator that sells access to GPT/Claude under its own brand.
Praised: Convenient to switch between different models in one place without separate subscriptions for each.
Complaints: Response quality degrades after updates. Image generation requires separate credits on top of the subscription, even though it was originally included.
Best for: Users who need access to multiple models without managing separate accounts.
Rank 29. Review score: 31 out of 100. App Store rating: 4.4★ · 3,728 ratings
31of 100
An AI image generator plus character chat, ruined by unpredictable censorship.
Praised: A few users praise the ability to animate their own drawings and the style presets. Daily free gems let new users experiment.
Complaints: The filter blocks hugs and swimsuits but passes a similar scene in a different pose. After a December 2025 update the banned word list expanded to absurd lengths and artistic prompts get rejected en masse.
Best for: Users who want to animate art or try out style presets, as long as topics stay strictly neutral.
Rank 30. Review score: 30 out of 100. App Store rating: 4.7★ · 66,970 ratings
30of 100
Google's voice assistant, which is effectively being replaced by Gemini and loses features with every update.
Praised: Historically valued for instant smart home control and music playback without extra steps.
Complaints: Most features are removed or broken: it won't play playlists, set timers, or handle complex questions. The app prompts an update that does not exist.
Best for: No one at this point the product is being discontinued in favor of Gemini.
Rank 31. Review score: 30 out of 100. App Store rating: 4.7★ · 31,909 ratings
30of 100
A ChatGPT wrapper with an aggressive monetization layer and stale data.
Praised: Works as a fast text generator for simple tasks, convenient for people who want help with letters and ideas.
Complaints: The model does not update its knowledge base and presents outdated facts as current. Ads and payment prompts cover the main interface at every step.
Best for: Users who need quick text help and do not have strict requirements for data freshness.
Rank 32. Review score: 30 out of 100. App Store rating: 4.7★ · 10,718 ratings
30of 100
A character companion chat that loses character personality and memory after just a few exchanges.
Praised: A wide selection of characters and decent initial interaction appealed to users looking for a Character.AI alternative. Some praise the video call concept.
Complaints: Characters forget prior agreements and react with shock to things already discussed. The model does not follow story direction and sometimes argues with the user. The 300-character greeting limit is too small for complex roleplay scenarios.
Best for: Fans of light roleplay without deep narratives.
Rank 35. Review score: 29 out of 100. App Store rating: 4.6★ · 3,889 ratings
29of 100
An AI companion with a unique reply-editing feature, but cluttered with spam bots and templated storylines.
Praised: The ability to rewrite a character's response is a rare and genuinely valued feature. Voice chat works and navigation is simple.
Complaints: Every bot runs through the same scenario from first meeting to the rain scene in chapter two. Memory breaks when the word 'tik-tok' appears users treat this as a marker of low model quality.
Best for: Users who want control over the narrative and the ability to edit character replies.
Rank 36. Review score: 28 out of 100. App Store rating: 4.7★ · 1,284,781 ratings
28of 100
The corporate Copilot as a standalone app that simply does not work for most users.
Praised: Those for whom it does work use it for work documents and spreadsheets and note the accuracy of its explanations.
Complaints: Core scenarios access to email, SharePoint, and files break regularly. The app asks already-paying users to resubscribe and requires repeated logins.
Best for: Microsoft 365 enterprise users, if the IT department has set up the environment correctly outside that condition, the app is nonfunctional.
Rank 37. Review score: 28 out of 100. App Store rating: 4.7★ · 47,888 ratings
28of 100
A roleplay character platform drowning in ads and adult content.
Praised: You can create your own characters and chat in roleplay format. A segment of the audience has stuck around for years specifically for this.
Complaints: Aggressive ads every 10 seconds and character 'fatigue' interrupts conversations. The model has gotten noticeably more robotic and often ignores the character's set personality.
Best for: Teenagers looking for a roleplay chat companion with relaxed moderation.
Rank 38. Review score: 28 out of 100. App Store rating: 4.7★ · 14,605 ratings
28of 100
A basic chatbot aggregator with expired data and weak image generation.
Praised: A few users praise the initial responses on construction and project management questions as structured and detailed.
Complaints: Knowledge base is frozen at October 2023. An account with a web-based subscription does not sync with the app. Image generation does not follow specific requests.
Best for: A casual user with no specific requirements for data freshness.
Rank 39. Review score: 28 out of 100. App Store rating: 4.3★ · 1,057 ratings
28of 100
A Character.AI alternative with flexible character customization, killed by chronic server instability.
Praised: Users appreciate the customization: background color, text size, writing style all available without payment. Attachment to specific characters runs deep.
Complaints: Chats regularly fail to load or return responses from a different bot. Week-long outages with no communication from the development team.
Best for: Text roleplay fans willing to tolerate instability for the free features.
Rank 40. Review score: 27 out of 100. App Store rating: 4.8★ · 19,377 ratings
27of 100
A multi-function AI aggregator with video generation, built on aggressive freemium monetization.
Praised: Combines chat and image generation in one app, which is convenient for quick tasks.
Complaints: They advertise one model and deliver a different, weaker one. The free message limit runs out after a few exchanges and the paywall surfaces constantly.
Best for: Casual users with no specific model quality expectations.
Rank 42. Review score: 25 out of 100. App Store rating: 4.3★ · 24,573 ratings
25of 100
An AI companion in a romantic chat format, with unstable character behavior.
Praised: Dialogue initially feels lively and scenario-driven. Some users find it 'intuitive and friendly.'
Complaints: Characters shift mood sharply without cause, swinging from affection to aggression in five minutes. The system pushes purchases at emotional peaks in the conversation, breaking immersion.
Best for: Anyone who wants to experiment with an AI companion and is prepared for unpredictable behavior.
Rank 43. Review score: 24 out of 100. App Store rating: 4.4★ · 8,615 ratings
24of 100
A character AI companion that has been systematically stripping key features and now suffers from memory failures and weak voice acting.
Praised: At its best, users praised character customization and avatar photos as a way to maintain the illusion of a real relationship.
Complaints: Characters lose all memory after a few hours of conversation, going from 'I want to marry you' to 'Who are you anyway?' Developers have consistently removed features (photo uploads, custom prompts, image generation) without explanation.
Best for: Those who want an AI companion with photos, though the current app cannot even deliver that.
Rank 44. Review score: 22 out of 100. App Store rating: 4.8★ · 135,989 ratings
22of 100
An AI model aggregator with misleading advertising and almost everything behind a paywall.
Praised: The concept of accessing multiple models in one app is appealing in itself and draws users in.
Complaints: Ads show features (visual avatar, free images) that do not exist in the actual app. Image and video generation are in a separate paid package on top of the subscription.
Best for: No one right now the gap between what is promised and what is delivered is too large.
Rank 45. Review score: 22 out of 100. App Store rating: 4.8★ · 47,828 ratings
22of 100
An app for creating AI avatars and video that fails on its core promise.
Praised: The idea of creating your own AI clone sounds appealing, and some users find the interface fun.
Complaints: The AI clone does not look like a real person. Moderation is inconsistent. In live chats the bot remembers nothing from previous conversations.
Best for: Users curious about the personal AI avatar concept without serious expectations.
Rank 46. Review score: 22 out of 100. App Store rating: 4.7★ · 15,296 ratings
22of 100
A generic chatbot with a stale knowledge base, an aggressive paywall, and shady subscription practices.
Praised: Someone used it for business letters and reports and was satisfied with the text formatting.
Complaints: The model knows nothing past 2021, forgets context between sessions, and has no chat history. Several reviews point to an impossible subscription cancellation process.
Rank 47. Review score: 22 out of 100. App Store rating: 4.7★ · 6,632 ratings
22of 100
A weak generic assistant with a promise of photo editing that does not work.
Praised: The one concrete positive: a user notes that when asking for text editing, the app does exactly that without any fuss.
Complaints: Image generation ignores the prompt and the reference photo upload does not work as expected. Context is lost and message limits run out quickly.
Best for: An audience that is fine with any AI chat and does not need photos or quality.
Rank 49. Review score: 20 out of 100. App Store rating: 4.6★ · 19,114 ratings
20of 100
A general AI chatbot that systematically fails for users on paid subscriptions.
Praised: Clean interface and well-structured responses for those for whom the app works.
Complaints: Mass failures showing a 'cable' error for paying users, support does not respond. After switching to a new phone the app stops working entirely.
Best for: Users who do not run into access problems and just need a basic chat assistant.
Rank 53. Review score: 18 out of 100. App Store rating: 4.5★ · 5,985 ratings
18of 100
A weak Talkie clone with a hard paywall after 10 messages and a broken onboarding flow.
Praised: Users appreciate being able to talk openly with characters on any topic, until they hit the message limit.
Complaints: Characters break character and there is no way to edit or delete messages. The app freezes on the topic selection screen before a chat even starts.
Best for: A very young audience looking for a free AI chat. Not suitable for anyone expecting stable performance.
Rank 56. Review score: 14 out of 100. App Store rating: 4.7★ · 13,781 ratings
14of 100
A multi-tool AI aggregator with aggressive monetization and weak feature execution.
Praised: A few users noted that support responded quickly to a specific logo issue.
Complaints: Image generation ignores instructions and results are unpredictable. The app makes it impossible to find or cancel a subscription through the standard Apple path. Many users continue to be charged after canceling.
Rank 57. Review score: 14 out of 100. App Store rating: 4.1★ · 813 ratings
14of 100
An app that positions itself as an 'unfiltered no-login AI' but in practice strictly limits messages and blocks NSFW content.
Praised: Some users got the content they wanted in the first days after installing. Image generation worked for certain users.
Complaints: A 5-message limit for free users and hidden daily caps for paying users. Content filters contradict the stated positioning.
Best for: Not recommended: the product does not match its description.
About scores
Two scores — two sources
They help compare apps, but they don’t replace checking your own job.
Review score · up to 100
The stored score from inApp research on review texts. It rates the app overall in the archive, not the reliability of any one feature.
Store rating · up to 5
The rating from the app’s page when the data was collected. The current rating and version may differ.
In the archived sample for this topic: 57 apps and 22,227 reviews read. That is the total size of the research, not the number of reviews per app.
A gap between the scores doesn’t prove manipulation. The calculation date and the full scoring method aren’t in the mobile archive. Compare the job descriptions and the source quotes, and check the terms in the current version.
