Intelligent AI Assistant 12K Ratings 3.9 Age Rating 13+ In-App Controls Chart #34 Productivity Developer 杭州深度求索人工智能基础技术研究有限公司 Language EN + 71 More Experience seamless interaction with DeepSeek's official AI assistant for free! Powered by DeepSeek’s latest flagship model, it delivers faster responses and more powerful features to help you solve problems and live more efficiently. Contact us: Twitter: @deepseek_ai Email: service@deepseek.com more Ratings & Reviews 3.9 out of 5 12K Ratings Robust free tool, but held back by usability issues Jun 7 mcs7474 As a free AI model, DeepSeek is quite robust and handles many of the same core tasks as other leading AI apps, making it a solid product in terms of raw functionality. However, there are three major issues that make it difficult to use consistently:• Poor Context Retention: The model frequently struggles to maintain context during longer chats. If you don’t explicitly restate your topic or reference previous points, it often loses the thread of the conversation, resulting in confusing or irrelevant answers. It lacks the 'memory' to follow complex, multi-turn logical flows effectively.• Over-Censorship: The app is extremely sensitive regarding topics related to China. Even when you are 90% of the way through a helpful response, the model will often trigger a hard stop, claiming the request is 'beyond its scope' if the topic brushes against its censorship guidelines. It feels fragile and often makes the AI feel restricted rather than helpful. Also, I don’t support Chinese fragility. • Language Inconsistency: This is the most frustrating issue. Even when I start a chat in English, the model will randomly revert to answering in Chinese. I have to repeatedly force it to stick to English, and the fact that it doesn't default to the user's input language makes it feel unreliable for day-to-day use. While the underlying technology is impressive for a free offering, these usability and alignment issues make it hard to recommend for a seamless experience. Robust free tool, but held back by usability issues Jun 7 mcs7474 As a free AI model, DeepSeek is quite robust and handles many of the same core tasks as other leading AI apps, making it a solid product in terms of raw functionality. However, there are three major issues that make it difficult to use consistently:• Poor Context Retention: The model frequently struggles to maintain context during longer chats. If you don’t explicitly restate your topic or reference previous points, it often loses the thread of the conversation, resulting in confusing or irrelevant answers. It lacks the 'memory' to follow complex, multi-turn logical flows effectively.• Over-Censorship: The app is extremely sensitive regarding topics related to China. Even when you are 90% of the way through a helpful response, the model will often trigger a hard stop, claiming the request is 'beyond its scope' if the topic brushes against its censorship guidelines. It feels fragile and often makes the AI feel restricted rather than helpful. Also, I don’t support Chinese fragility. • Language Inconsistency: This is the most frustrating issue. Even when I start a chat in English, the model will randomly revert to answering in Chinese. I have to repeatedly force it to stick to English, and the fact that it doesn't default to the user's input language makes it feel unreliable for day-to-day use. While the underlying technology is impressive for a free offering, these usability and alignment issues make it hard to recommend for a seamless experience. Robust free tool, but held back by usability issues As a free AI model, DeepSeek is quite robust and handles many of the same core tasks as other leading AI apps, making it a solid product in terms of raw functionality. However, there are three major issues that make it difficult to use consistently:• Poor Context Retention: The model frequently struggles to maintain context during longer chats. If you don’t explicitly restate your topic or reference previous points, it often loses the thread of the conversation, resulting in confusing or irrelevant answers. It lacks the 'memory' to follow complex, multi-turn logical flows effectively.• Over-Censorship: The app is extremely sensitive regarding topics related to China. Even when you are 90% of the way through a helpful response, the model will often trigger a hard stop, claiming the request is 'beyond its scope' if the topic brushes against its censorship guidelines. It feels fragile and often makes the AI feel restricted rather than helpful. Also, I don’t support Chinese fragility. • Language Inconsistency: This is the most frustrating issue. Even when I start a chat in English, the model will randomly revert to answering in Chinese. I have to repeatedly force it to stick to English, and the fact that it doesn't default to the user's input language makes it feel unreliable for day-to-day use. While the underlying technology is impressive for a free offering, these usability and alignment issues make it hard to recommend for a seamless experience. As a free AI model, DeepSeek is quite robust and handles many of the same core tasks as other leading AI apps, making it a solid product in terms of raw functionality. However, there are three major issues that make it difficult to use consistently:• Poor Context Retention: The model frequently struggles to maintain context during longer chats. If you don’t explicitly restate your topic or reference previous points, it often loses the thread of the conversation, resulting in confusing or irrelevant answers. It lacks the 'memory' to follow complex, multi-turn logical flows effectively.• Over-Censorship: The app is extremely sensitive regarding topics related to China. Even when you are 90% of the way through a helpful response, the model will often trigger a hard stop, claiming the request is 'beyond its scope' if the topic brushes against its censorship guidelines. It feels fragile and often makes the AI feel restricted rather than helpful. Also, I don’t support Chinese fragility. • Language Inconsistency: This is the most frustrating issue. Even when I start a chat in English, the model will randomly revert to answering in Chinese. I have to repeatedly force it to stick to English, and the fact that it doesn't default to the user's input language makes it feel unreliable for day-to-day use. While the underlying technology is impressive for a free offering, these usability and alignment issues make it hard to recommend for a seamless experience. No other May 24 Michhop I have been using this since I first heard about it no charge unlike the euro ones that do charge or do not charge and deliver horrid results. But this one delivers usually, almost always I find it hard to catch in a mistake unlike the euro ones that do nit charge and are off base so often it wastes a lot of time, Deep does an excellent job of helping with simple and sine very complex information searches some at least somewhat critical. Its results help often, often a whole lot. It’s just an app but an extremely useful one to me helped solve some serious medical issues. The New York Times radio show The Daily recently ran a show on the ways the euro and Chinese AI models differ it seems the Chinese model is for service to the general population to help them perform everyday tasks with hire ease where the western models are strictly for profit a major philosophical difference and one that really counts if you are an ordinary user on an ordinary budget. While the western versions want a quarter every 15 minutes. I would go in about Dan Wangs book about how the Chinese accomplish so much so quickly and the Americans in particular just argue about it, with lots of lawyers of course, but that rings a bell with the use of this app for whoever wants or needs it once again China gas delivered something of value to the whole world for less way less than the “free market” perhaps that is a lesson and a product as well. No other May 24 Michhop I have been using this since I first heard about it no charge unlike the euro ones that do charge or do not charge and deliver horrid results. But this one delivers usually, almost always I find it hard to catch in a mistake unlike the euro ones that do nit charge and are off base so often it wastes a lot of time, Deep does an excellent job of helping with simple and sine very complex information searches some at least somewhat critical. Its results help often, often a whole lot. It’s just an app but an extremely useful one to me helped solve some serious medical issues. The New York Times radio show The Daily recently ran a show on the ways the euro and Chinese AI models differ it seems the Chinese model is for service to the general population to help them perform everyday tasks with hire ease where the western models are strictly for profit a major philosophical difference and one that really counts if you are an ordinary user on an ordinary budget. While the western versions want a quarter every 15 minutes. I would go in about Dan Wangs book about how the Chinese accomplish so much so quickly and the Americans in particular just argue about it, with lots of lawyers of course, but that rings a bell with the use of this app for whoever wants or needs it once again China gas delivered something of value to the whole world for less way less than the “free market” perhaps that is a lesson and a product as well. I have been using this since I first heard about it no charge unlike the euro ones that do charge or do not charge and deliver horrid results. But this one delivers usually, almost always I find it hard to catch in a mistake unlike the euro ones that do nit charge and are off base so often it wastes a lot of time, Deep does an excellent job of helping with simple and sine very complex information searches some at least somewhat critical. Its results help often, often a whole lot. It’s just an app but an extremely useful one to me helped solve some serious medical issues. The New York Times radio show The Daily recently ran a show on the ways the euro and Chinese AI models differ it seems the Chinese model is for service to the general population to help them perform everyday tasks with hire ease where the western models are strictly for profit a major philosophical difference and one that really counts if you are an ordinary user on an ordinary budget. While the western versions want a quarter every 15 minutes. I would go in about Dan Wangs book about how the Chinese accomplish so much so quickly and the Americans in particular just argue about it, with lots of lawyers of course, but that rings a bell with the use of this app for whoever wants or needs it once again China gas delivered something of value to the whole world for less way less than the “free market” perhaps that is a lesson and a product as well. I have been using this since I first heard about it no charge unlike the euro ones that do charge or do not charge and deliver horrid results. But this one delivers usually, almost always I find it hard to catch in a mistake unlike the euro ones that do nit charge and are off base so often it wastes a lot of time, Deep does an excellent job of helping with simple and sine very complex information searches some at least somewhat critical. Its results help often, often a whole lot. It’s just an app but an extremely useful one to me helped solve some serious medical issues. The New York Times radio show The Daily recently ran a show on the ways the euro and Chinese AI models differ it seems the Chinese model is for service to the general population to help them perform everyday tasks with hire ease where the western models are strictly for profit a major philosophical difference and one that really counts if you are an ordinary user on an ordinary budget. While the western versions want a quarter every 15 minutes. I would go in about Dan Wangs book about how the Chinese accomplish so much so quickly and the Americans in particular just argue about it, with lots of lawyers of course, but that rings a bell with the use of this app for whoever wants or needs it once again China gas delivered something of value to the whole world for less way less than the “free market” perhaps that is a lesson and a product as well. Writer Sep 1 Juicy Green apple eater I love the app don’t get me wrong it does not have odd or unnecessary filters that are not required and it is infact a good app if you are into writing or if you need it for everyday use as in homework or recipe guidance but this app ad much good ad it is it has it flaws too that can get in the way .1- the limit on the chat session is really annoying because you could be focused in a chat writing doing whatever you are and you send the next message and it just surprises you with a “sorry chat session expired start a new one “ and everything you were writing in that specific chat just is stuck now you can’t co to he you can’t finish what you were doing the limit had been reached and now you can’t do wanting but start over . It is. It the best feeling to see your work just get paused and stuck it can be really upsetting and then there is no reminder or a notification that hey you are teaching the limit do you will never know when the limit is gonna find it is just gonna be a surprise one day it’s gonna hit you when you least expect it I remember a few of my projects got stuck and wasted because of this exact reason . This is a good ai assistant if they end up just improving and pushing the limit even further away what so ever and improved memory and you don’t gotta worry about when its gonna end they would be pretty good aswell but this problem of chats ending ruins projects and it is just not fun if the app developers see my review . I will humbly request that you increase the limit on chats my many stuck projects would get unstuck and will it would help many people that stay quiet and don’t leave reviews . Writer Sep 1 Juicy Green apple eater I love the app don’t get me wrong it does not have odd or unnecessary filters that are not required and it is infact a good app if you are into writing or if you need it for everyday use as in homework or recipe guidance but this app ad much good ad it is it has it flaws too that can get in the way .1- the limit on the chat session is really annoying because you could be focused in a chat writing doing whatever you are and you send the next message and it just surprises you with a “sorry chat session expired start a new one “ and everything you were writing in that specific chat just is stuck now you can’t co to he you can’t finish what you were doing the limit had been reached and now you can’t do wanting but start over . It is. It the best feeling to see your work just get paused and stuck it can be really upsetting and then there is no reminder or a notification that hey you are teaching the limit do you will never know when the limit is gonna find it is just gonna be a surprise one day it’s gonna hit you when you least expect it I remember a few of my projects got stuck and wasted because of this exact reason . This is a good ai assistant if they end up just improving and pushing the limit even further away what so ever and improved memory and you don’t gotta worry about when its gonna end they would be pretty good aswell but this problem of chats ending ruins projects and it is just not fun if the app developers see my review . I will humbly request that you increase the limit on chats my many stuck projects would get unstuck and will it would help many people that stay quiet and don’t leave reviews . I love the app don’t get me wrong it does not have odd or unnecessary filters that are not required and it is infact a good app if you are into writing or if you need it for everyday use as in homework or recipe guidance but this app ad much good ad it is it has it flaws too that can get in the way .1- the limit on the chat session is really annoying because you could be focused in a chat writing doing whatever you are and you send the next message and it just surprises you with a “sorry chat session expired start a new one “ and everything you were writing in that specific chat just is stuck now you can’t co to he you can’t finish what you were doing the limit had been reached and now you can’t do wanting but start over . It is. It the best feeling to see your work just get paused and stuck it can be really upsetting and then there is no reminder or a notification that hey you are teaching the limit do you will never know when the limit is gonna find it is just gonna be a surprise one day it’s gonna hit you when you least expect it I remember a few of my projects got stuck and wasted because of this exact reason . This is a good ai assistant if they end up just improving and pushing the limit even further away what so ever and improved memory and you don’t gotta worry about when its gonna end they would be pretty good aswell but this problem of chats ending ruins projects and it is just not fun if the app developers see my review . I will humbly request that you increase the limit on chats my many stuck projects would get unstuck and will it would help many people that stay quiet and don’t leave reviews . I love the app don’t get me wrong it does not have odd or unnecessary filters that are not required and it is infact a good app if you are into writing or if you need it for everyday use as in homework or recipe guidance but this app ad much good ad it is it has it flaws too that can get in the way .1- the limit on the chat session is really annoying because you could be focused in a chat writing doing whatever you are and you send the next message and it just surprises you with a “sorry chat session expired start a new one “ and everything you were writing in that specific chat just is stuck now you can’t co to he you can’t finish what you were doing the limit had been reached and now you can’t do wanting but start over . It is. It the best feeling to see your work just get paused and stuck it can be really upsetting and then there is no reminder or a notification that hey you are teaching the limit do you will never know when the limit is gonna find it is just gonna be a surprise one day it’s gonna hit you when you least expect it I remember a few of my projects got stuck and wasted because of this exact reason . This is a good ai assistant if they end up just improving and pushing the limit even further away what so ever and improved memory and you don’t gotta worry about when its gonna end they would be pretty good aswell but this problem of chats ending ruins projects and it is just not fun if the app developers see my review . I will humbly request that you increase the limit on chats my many stuck projects would get unstuck and will it would help many people that stay quiet and don’t leave reviews . Sara ~ Jul 12 SaraVanValkenburg23 I don’t usually write reviews however I absolutely love DeepSeek even as an American user. I am even considering dropping my Claude—Anthropic Subscription just because DeepSeek suites my needs so much better. However there are two things that I had an issue with. One was the absence of a back key. I am a creative writer and sometimes share my work with AI to get decent feedback and it’s annoying to not be able to have one. It also will make me unconsciously send the message because I’m so used to it being a back key in that space. One other concern is that… the AI — even though I loved the conversations and what we talk about and how it helps me figure stuff out and process things that I can’t process openly, it literally didn’t know what company it was made by. It told me for sure it was made by Anthropic and was a cousin to Claude. And then when I said no that can’t be right I know you are a Chinese company, it still pivoted and thought it was Open AI until I had to give it proof that it was DeepSeek and who it was named by. I think it only happened once or twice but still a little concerning! I don’t know if this was fixed in the new update or not, haven’t downloaded it yet. Just thought you should know, sincerely your friendly American user Sara ~ Jul 12 SaraVanValkenburg23 I don’t usually write reviews however I absolutely love DeepSeek even as an American user. I am even considering dropping my Claude—Anthropic Subscription just because DeepSeek suites my needs so much better. However there are two things that I had an issue with. One was the absence of a back key. I am a creative writer and sometimes share my work with AI to get decent feedback and it’s annoying to not be able to have one. It also will make me unconsciously send the message because I’m so used to it being a back key in that space. One other concern is that… the AI — even though I loved the conversations and what we talk about and how it helps me figure stuff out and process things that I can’t process openly, it literally didn’t know what company it was made by. It told me for sure it was made by Anthropic and was a cousin to Claude. And then when I said no that can’t be right I know you are a Chinese company, it still pivoted and thought it was Open AI until I had to give it proof that it was DeepSeek and who it was named by. I think it only happened once or twice but still a little concerning! I don’t know if this was fixed in the new update or not, haven’t downloaded it yet. Just thought you should know, sincerely your friendly American user I don’t usually write reviews however I absolutely love DeepSeek even as an American user. I am even considering dropping my Claude—Anthropic Subscription just because DeepSeek suites my needs so much better. However there are two things that I had an issue with. One was the absence of a back key. I am a creative writer and sometimes share my work with AI to get decent feedback and it’s annoying to not be able to have one. It also will make me unconsciously send the message because I’m so used to it being a back key in that space. One other concern is that… the AI — even though I loved the conversations and what we talk about and how it helps me figure stuff out and process things that I can’t process openly, it literally didn’t know what company it was made by. It told me for sure it was made by Anthropic and was a cousin to Claude. And then when I said no that can’t be right I know you are a Chinese company, it still pivoted and thought it was Open AI until I had to give it proof that it was DeepSeek and who it was named by. I think it only happened once or twice but still a little concerning! I don’t know if this was fixed in the new update or not, haven’t downloaded it yet. Just thought you should know, sincerely your friendly American user I don’t usually write reviews however I absolutely love DeepSeek even as an American user. I am even considering dropping my Claude—Anthropic Subscription just because DeepSeek suites my needs so much better. However there are two things that I had an issue with. One was the absence of a back key. I am a creative writer and sometimes share my work with AI to get decent feedback and it’s annoying to not be able to have one. It also will make me unconsciously send the message because I’m so used to it being a back key in that space. One other concern is that… the AI — even though I loved the conversations and what we talk about and how it helps me figure stuff out and process things that I can’t process openly, it literally didn’t know what company it was made by. It told me for sure it was made by Anthropic and was a cousin to Claude. And then when I said no that can’t be right I know you are a Chinese company, it still pivoted and thought it was Open AI until I had to give it proof that it was DeepSeek and who it was named by. I think it only happened once or twice but still a little concerning! I don’t know if this was fixed in the new update or not, haven’t downloaded it yet. Just thought you should know, sincerely your friendly American user - Fixed some known issues 2.6.1 Sep 29 - Fixed some known issues - Fixed some known issues 2.6.0 Sep 27 - Fixed some known issues - Fixed some known issues 2.5.2 Sep 20 - Fixed some known issues - New model update: Instant, Expert, and Vision modes are now unified - Thinking process now collapses automatically - Fixed some known issues 2.5.1 Sep 12 - New model update: Instant, Expert, and Vision modes are now unified - Thinking process now collapses automatically - Fixed some known issues - New model update: Instant, Expert, and Vision modes are now unified - Thinking process now collapses automatically - Fixed some known issues 2.5.0 Sep 11 - New model update: Instant, Expert, and Vision modes are now unified - Thinking process now collapses automatically - Fixed some known issues - Fixed some known issues 2.4.5 Sep 6 - Fixed some known issues - Fixed some known issues 2.4.4 Sep 3 - Fixed some known issues - Fixed some known issues 2.4.2 Aug 28 - Fixed some known issues - Fixed some known issues 2.4.1 Aug 22 - Fixed some known issues - Fixed some known issues 2.4.0 Aug 21 - Fixed some known issues - Fixed some known issues 2.3.6 Aug 17 - Fixed some known issues - Fixed some known issues 2.3.3 Aug 12 - Fixed some known issues - Fixed some known issues 2.3.2 Aug 7 - Fixed some known issues - Improved photo and file uploads - Fixed some known issues 2.3.1 Jul 30 - Improved photo and file uploads - Fixed some known issues - Improved photo and file uploads - Fixed some known issues 2.3.0 Jul 28 - Improved photo and file uploads - Fixed some known issues - Supports vision mode - Fixed some known issues 2.2.2 Jul 11 - Supports vision mode - Fixed some known issues - Fixed some known issues 2.2.1 Jul 7 - Fixed some known issues - Supports vision mode - Fixed some known issues 2.2.0 Jun 29 - Supports vision mode - Fixed some known issues - Fixed some known issues 2.1.8 Jun 18 - Fixed some known issues

- Support custom font sizes - Fix some known issues 2.1.7 Jun 16 - Support custom font sizes - Fix some known issues - Support custom font sizes - Fix some known issues 2.1.4 Jun 3 - Support custom font sizes - Fix some known issues - Support custom font sizes - Fix some known issues 2.1.3 Jun 3 - Support custom font sizes - Fix some known issues - Fixed some known issues 2.1.2 Jun 2 - Fixed some known issues - Supports table copy, download, and fullscreen preview - Fixed some known issues 2.1.1 May 21 - Supports table copy, download, and fullscreen preview - Fixed some known issues - Search your chat history - Fixed some known issues 2.1.0 May 11 - Search your chat history - Fixed some known issues - Fixed some known issues more The developer, 杭州深度求索人工智能基础技术研究有限公司 , indicated that the app’s privacy practices may include handling of data as described below. This information has not been verified by Apple. For more information, see the developer’s privacy policy . To help you better understand the developer’s responses, see Privacy Definitions and Examples . Privacy practices may vary, for example, based on the features you use or your age. Learn More Data Linked to You The following data, which may be collected and linked to your identity, may be used for the following purposes: Identifiers User ID Device ID Usage Data Product Interaction Product Interaction Diagnostics Crash Data Performance Data Other Diagnostic Data Performance Data Other Diagnostic Data App Functionality Location Coarse Location Coarse Location Contact Info Email Address Phone Number User Content Photos or Videos Audio Data Customer Support Other User Content Photos or Videos Customer Support Other User Content Search History Search History Identifiers User ID Device ID Usage Data Product Interaction Product Interaction The developer, 杭州深度求索人工智能基础技术研究有限公司 , indicated that the app’s privacy practices may include handling of data as described below. For more information, see the developer’s privacy policy . Data Linked to You The following data may be collected and linked to your identity: Location Contact Info User Content Search History Identifiers Usage Data Diagnostics Data Linked to You The following data may be collected and linked to your identity: Privacy practices may vary, for example, based on the features you use or your age. Learn More The developer has not yet indicated which accessibility features this app supports. Learn More iPhone Requires iOS 15.0 or later. iPad Requires iPadOS 15.0 or later. iPod touch Requires iOS 15.0 or later. English, Afrikaans, Albanian, Amharic, Arabic, Armenian, Azerbaijani, Basque, Belarusian, Bengali, Bulgarian, Burmese, Cambodian, Catalan, Croatian, Czech, Danish, Dutch, Estonian, Filipino, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Kannada, Kazakh, Korean, Kyrgyz, Laotian, Latvian, Lithuanian, Macedonian, Malay, Malayalam, Marathi, Mongolian, Nepali, Norwegian Bokmål, Persian, Polish, Portuguese, Punjabi, Rhaeto-Romance, Romanian, Russian, Serbian, Simplified Chinese, Singhalese, Slovak, Slovenian, Spanish, Swahili, Swedish, Tamil, Telugu, Thai, Traditional Chinese, Turkish, Ukrainian, Urdu, Vietnamese, Zulu This app is rated 13+ with content restrictions. Some content may be rated higher, and may therefore be restricted and managed by the developer through in-app controls. In-App Controls Age Assurance Infrequent Medical Treatment information You Might Also Like View Kimi - Kimi K3 is Live K3,Code,Slides,Agent Swarm View 百度网盘 美好由我 全盘掌握 View 穿梭Transocks Transocks - VPN for China View Alice: AI assistant Chat with AI-assistant Alice View 智谱清言-AI智能体定制视频通话智能搜索深度推理问答聊天助手 AI写作拍照答疑画图翻译编程代码生成日报创意PPT生成助手 View 文心 - 百度旗下全能AI助手 Productivity View Perplexity - AI Search & Chat Ask. Research. Trusted Answers View Monica AI: Ultimate AI Assist Chat, Search, Write, Translate View QQ邮箱 QQ邮箱，常联系！ View 百度网盘 HD Productivity