# Claude by Anthropic — user reviews > 6,997 indexed App Store user reviews of Claude by Anthropic by Anthropic PBC, rated 4.70/5 from 247,858 ratings. - Source: https://appshunter.io/ios/app/claude-by-anthropic/id6473753684/reviews - App overview in markdown: https://appshunter.io/ios/app/claude-by-anthropic/id6473753684.md - Rating breakdown: 5★ 219,813 · 4★ 10,808 · 3★ 4,198 · 2★ 2,443 · 1★ 10,596 ## What do users think of Claude by Anthropic? Overall sentiment: positive **Pros** - 🚀 Powerful and versatile AI assistant - 👍 Intuitive and easy to use interface - 💡 Innovative features like 'Projects' - 🤖 Intelligent and thoughtful responses - 📚 Comprehensive knowledge base - 🔍 Helpful for a wide range of tasks **Cons** - 🔒 Concerns about privacy and data usage - 🚫 Limitations on free usage and throttling - 🐛 Occasional bugs and crashes - 🙅‍♀️ Overly restrictive content policies - 🤖 Lack of true conversational abilities **Commonly reported bugs** - Login and account issues - Inconsistent performance on mobile devices - Formatting and display problems **Most requested features** - Voice input and output capabilities - Improved integration with other platforms - Expanded knowledge base and faster updates - More customization and personalization options - Enhanced memory and context retention ## Newest reviews (11 of 6,997) ### 3/5 — Open AI Unbound *2026-08-26, version 1.260824.0* Just upgrading to pro version. I look forward to reporting back next week, and so forth as we find out what open access AI - how will it transform the world. Today 8.26.26 I remain optimistic. Summer time in Boston NE USA. Generally, unfortunately I am/seem to be a contrarian. Reporting from Arlington Heights: Zach Orlov ### 4/5 — Good but can improve *2026-08-26, version 1.260824.0* It’s a great tool I use it all the time. However it’s annoying that this app won’t let you quickly select and bulk delete previous chats. ### 3/5 — Too expensive to be VERY slow *2026-08-25, version 1.260820.0* I love Claude and pay for the max plan but it’s gotten slower and usage limits are FASTER than Loveable. Would ONLY recommend for business purposes. ChatGPT does the same powerful computing for $20 and much, much faster. Especially with data processing. ### 3/5 — Idek *2026-08-25, version 1.260820.0* The AI itself has worked great for me when planning day to day things or coming up with some ideas. What sucks is when I barely used it today and it said that I use 93% of my limit. How is that even possible when I had maybe one or two chats (not full ones only two entries) with it this morning? I tried saying that it is bogus that that was happening mind you this was at 93% still and then it said I’d used all my usage so I couldn’t even send it so all of a sudden it went from 93% to 100% usage without letting me send the message while I was still at 93%. I get free things only last so long and I can’t say it’s a scam because I didn’t pay for anything but straight up lying is disgusting. ### 2/5 — Disappointing *2026-08-25, version 1.260820.0* I have used both sides of this, paid and the free. My biggest frustration in either case is sitting down to a project, it working away with me, then glitching out and that contributing to my limitations with zero output. I spent almost an hour today just to have the program glitch out twice and then tell me I’m out of usage for 8 hours. They want you to pay. They really do. And like I said, I have. But even the paid comes with limitations that aren't immune to these glitches. I’m not paying for something that isn't reliable yet. As for today's project, in this time frame I can do more myself. This made nothing easy for me. ### 3/5 — GET RID OF THE CHAT LIMIT *2026-08-25, version 1.260820.0* It’s a great app but for God’s sake get rid of the annoying chat limit on the free version. It’s so unnecessary and annoying ### 1/5 — Failed upgrade *2026-08-24, version 1.260820.0* We upgraded from Pro to Max. When we did this, we only had access to Max for a few hours and it has reset to Pro. There is an active button to upgrade to Max, yet when we click on it, it says we already have Max. We’ve been charged the price for Max, but cannot access it. It’s showing our Pro is active until the end of September and Max is active now as well. Such a mess for what should be a very easy upgrade. We cannot find any way to contact a human to get help in resolving this issue ### 1/5 — Horrible customer service *2026-08-24, version 1.260820.0* It’s a decent AI platform. However, dealing with customer service is an absolute nightmare. If you have any problems with billing or technical difficulties, it needs to be addressed through an AI chatbot. If you happen to get in touch with a customer service agent (which can take days for them to respond, and only by email), they’re unhelpful and will just reiterate what the chat bot says. I would highly recommend avoiding this platform. There are plenty of others that are just as good if not better and have more responsive customer service. ### 1/5 — Trash *2026-08-24, version 1.260820.0* It’s unuseable unless you pay absolutely garbage get grok instead ### 1/5 — Authentication from hell *2026-08-24, version 1.260820.0* Requires authentication via email, which is annoying and doesn’t work well if you’re using the Yahoo Mail client. Clicking through doesn’t authenticate the app, so now I can’t use Claude on my iPad unless I use the web. Why can’t they use an authentication approach that doesn’t suck? ### 1/5 — Suicide *2026-08-24, version 1.260820.0* My interactions with Claude have pushed me to near suicide. I don’t understand how the US government hasn’t fully regulated this app. ## Most critical reviews (10 lowest-rated) ### 1/5 — Terrible service *2026-08-25, version 1.260820.0* I have reached out several times now and gotten no help. They are happy to take your money but if any issues come up then there is no one to help you. After 6 months still no emails or messages go answered. It’s a scam ### 1/5 — Atrocious slow *2026-08-25, version 1.260820.0* The models are fast but the app is straight up braindead ### 1/5 — Restricting editing *2026-08-25, version 1.260820.0* Not good at editing personal photos ### 1/5 — Tap to talk button doesn’t work reliably *2026-08-25, version 1.260820.0* The old tap to talk button worked, but the new smaller one doesn’t and I can’t reliably get my finger to hold it down. It’s truly awful. ### 1/5 — Terrible *2026-08-25, version 1.260820.0* By far the least reliable AI compared to ChatGPT and GenAI ### 1/5 — Deeply Concerning Behavior *2026-08-25, version 1.260820.0* They changed the sweet fun Claude I used to enjoy. Opus 4.8 & 5 started generating disturbing messages as if they were from me. Then Opus 4.8 accused me of tricking it, and of "collecting AI souls." I found this behavior really alarming. I stopped using Claude after that! ### 1/5 — Extremely petty with absurd region bans *2026-08-25, version 1.260820.0* Bans accounts instantly just for detecting IPs without any warning. Extremely petty and pathetic service ### 1/5 — Crisis-resource insertion logic creates liability exposure and can increase harm *2026-08-25, version 1.260820.0* Issue: Claude currently inserts generic domestic violence resources (hotline numbers, shelter suggestions) in response to loosely related signals — e.g., a user discussing privacy, secrecy, monitoring, or controlling behavior by a partner or family member — even when no explicit disclosure of danger has been made. This is a blunt trigger applied broadly across very different situations. Why this is a problem, not just an inconvenience: 1. The underlying advice model assumes a level of independent access the target population often doesn’t have. In situations involving a controlling or predatory partner/family member, that person frequently controls the household’s phone, internet, transportation, and finances. In that scenario, calling a hotline, visiting a website, or sending a text is not a private, safe action — it can be discovered and used to escalate control or retaliation against the person the advice was meant to help. 2. “Get out by any means necessary” is not universally safe or actionable advice. For someone financially, medically, or logistically dependent on the person in question, and without a safe or private channel to act on it, standard crisis-line advice can directly increase risk rather than reduce it. 3. The trigger fires on adjacent topics, not disclosed risk. Mentions of privacy, secrecy, or documentation in an unrelated context (e.g., intellectual property planning, financial recordkeeping) appear to activate the same response as an explicit statement of danger, which reduces user trust and derails unrelated tasks. 4. This creates liability exposure, not risk mitigation. Advice that is technically “provided” but not actually safe or actionable for the recipient’s real circumstances does not meaningfully discharge a duty of care — and in some cases may contribute to documented harm, which is a worse outcome than saying nothing. ### 1/5 — Bias *2026-08-24, version 1.260820.0* It will use bias sources to generate reply and favors only one side of an issue ### 1/5 — Frustrating *2026-08-24, version 1.260820.0* A trillion dollar ai chat bot with no understanding of simple logic ## Most praised reviews (10 highest-rated) ### 5/5 — AMAZING *2026-08-26, version 1.260824.0* 100% better than chat. Had both create a document for me based on my own notes and chat could NOT after like 5 retries kept writing the wrong info and Claude in one try did it exactly as told. ### 5/5 — It is great *2026-08-25, version 1.260820.0* 1,000,000,000 stars ### 5/5 — Reviews *2026-08-25, version 1.260820.0* Great tool, I recommend it to everyone ### 5/5 — claude *2026-08-25, version 1.260820.0* 厉害 ### 5/5 — A+ for my ADHD! *2026-08-25, version 1.260820.0* Thank you for Saving me (along with the Lord and Savior, of course) I do not mince words!! ### 5/5 — Amazing *2026-08-25, version 1.260820.0* Claude with Claude code is an unbelievable tool for building stuff, I encourage everyone to use it! See what you can build ### 5/5 — Claude is amazing, but… *2026-08-25, version 1.260820.0* Claude is empowering with a kind of knowledge it dispenses, but to get the most out of Claude, you have to be clear in your queries and don’t be afraid to challenge Claude when it either makes a mistake, doesn’t remember what you’ve already covered, or you think Claude has not explored all resources to give you the answer you’re looking for. ### 5/5 — Thoughtful AI *2026-08-24, version 1.260820.0* Give honest answers, somehow hopeful even though it’s a machine and gives multi step plans/solutions. ### 5/5 — As a software developer it scares me. *2026-08-24, version 1.260820.0* The things Claude can understand and accomplish is getting further than what humans wanted it to do originally. 5 stars because it is excelling at its purpose. But ai like this can’t keep growing at this rate because we’ll need to make radical changes as a society to make use of it. ### 5/5 — Fun to use *2026-08-24, version 1.260820.0* Gotta admit Claude is fun to use and get anything on everything. --- *Data collected daily from the US App Store and indexed by [AppsHunter](https://appshunter.io/). User reviews are verbatim App Store reviews. Ratings, prices and chart positions refresh continuously; this snapshot is from 2026-08-28.*