101 AI Apps You Can Vibe Code

Here are 101 AI apps that you can start building today and you don't even need to know how to code. Just as a note, I'm not saying that coding is obsolete. It is still very much necessary and useful for more complex and custom apps and features. But it is very much possible now to build certain types of applications without traditional coding aka bip coding. I have done many myself. So I'm going to split this video into seven different categories. Starting off with database/data management type apps, hardwarebased apps, dashboards, chatbot/ass assistant agent type apps, coaches/arning agent type apps, multimodality, and finally automation/mros. Now, before we dive into all the different app ideas, I do want to do a quick crash course on my five-step framework for how to approach building AI applications all the way from ideiation to deployment. This will give you a practical starting point for how to build these projects and the tools to use for this when you actually get really excited to build these applications. A portion of this video is sponsored by Bault. The five-step

framework is metaprompt, product requirements prompt, the PRP, implementation, debugging, and deployment. The first part of the framework is the meta prompt. And this is what I like to use in the very beginning in the ideation phase because it goes through and asks you questions like what's the purpose of the app? Who's the target audience? Who are you marketing? What are the core features? And it goes through all of these and forces you to think about who it is that your application is actually designed for. Because after all, if you don't even know exactly what it is that you're building, it's kind of unfair to expect your AI to figure out how to build it, right? So, I'm actually going to provide you the meta prompt, which I'll link in the description that I like to personally use. Okay. So, after you think very deeply and answer all of these questions, the output of this meta prompt is going to be called a product requirements prompt, which is the second part of this framework, a PRP. If you guys have heard about a product requirements document before, it's generally what a lot of product managers will use to have a very clear description of what their product is and all the core features surrounding it. So that meta prompt will help you generate this PRP. And what you do is that you take this PRP and you input it into your AI assisted coding tool. Just with a single prompt, you'll probably be able to get like 80 to 90% of your core set of features, which is honestly really impressive. But to get the rest of the 10 20% to actually make it into a full application, we do need to go on to the next third step which is the step of incremental implementation. You see, for the things that it doesn't get quite right, you do need to incrementally start asking it to change certain things like add a little bit more of this or change this, change the UI a little bit more. Um, you know, add something else like maybe add like an audio functionality. And actually step three and step four go hand in hand and that is debugging. as you're working with the AI and getting it to implement more things. You're also going to be coming across errors or things that the AI is just not doing correctly. This is really normal and part of the process. So when you do encounter these errors, the first thing that you want to do is just try to ask AI to fix it itself. Literally just be like there's usually a button that you can just click and be like, you know, try to fix the error itself. Um or you can do things like take a screenshot of the error or just point out that something is not working like this button is not working. Then the AI is usually pretty good at going back and then fixing these errors. This is also going to be incremental. So you're going to be repeating this process of incrementally implementing a feature and then debugging it, implementing, debugging, implementing, debugging until towards the end, you get to a point in which you're happy with the features that have been implemented and it's functioning properly. Then at that point, you're ready for step five, which is deployment. Honestly, there are a lot of things that I can talk about in this deployment section, including things like security, risk, version control, or to be hosting things. But the good news is that for most of these VIP coding AI assisted coding tools, they do have their own deployment options. Okay, so that is a really quick crash course on building an AI application all the way from ideation to the deployment stage. There is so much more that I can talk about and I actually made like full videos going through this process which I will link above over here and also in the description if you do want to go into more detail about these step-by-step processes. But for the rest of this video, I do want to focus on the AI apps that we are very excited to start building. Starting off with database/

data management type apps. This category of AI applications is all about leveraging AI's incredible ability of turning raw or unstructured data into something that you can search, query, summarize, and extract value from. The general workflow for this type of application usually starts with ingesting a lot of different types of information, labeling it, indexing it, and then organizing it into a nice little database. then allow users to be able to ask questions and the AI would go and filter all the information in the database to collect the relevant information maybe do some types of analysis if it has to and finally output it in some format maybe like visualizations an email or a report for example if you're building something like a video search database and a user can come in and ask I want to get the specific clip of this specific math question that was being shown by the professor uh through like catalog of hours and hours of lectures please please somebody built this by the way the workflow what it would look like is that you would first give it like a batch of different lectures all the different lectures that are available. The AI would like transcribe that information, restructure it, index it, and put it into a database. The user would be able to ask a question like find the specific formula from all of these different videos. The AI would search through the video content, extract that clip, and then show it to the user where they can like, I don't know, download it or export it, whatever, right? This kind of database search application is also really powerful for other types of modalities like images, quotes, text, and of course, just combinations of these different things as well. You can also make a cross tool enterprise search where an AI is going to be connected to the database sources from like notion and Slack and Google Drive and a lot of different things and you can unify all of that as a search engine. Audio transcription and search is a really big one if you have a lot of podcasts and a lot of meetings that you have there and you just want to pick out a very specific section and also semantic search over internal documents being able to search very specific things related to an internal company system. Sometimes the emphasis of an app can be on the transformation of that data, allowing the user to upload a bunch of PDFs and notes and then AI being able to organize that information um in terms of topics or Q& A sections. Data cleaning assistant. Maybe you have a bunch of Excel files that have like just columns and all over the place like everything is all over the place. There's missing values. Everything's not transformed correctly. You can build an AI app that's able to clean everything up and just allow the user to redownload the clean version. Image tagging or classification. Maybe you have like lots of different pictures about flowers and you want to upload that and then allow the AI to classify it into different types of flowers. Works for all different types of modalities. You can also create an app that annotates all the files that are being uploaded. Like if it sees a picture of a butterfly, it would actually put an annotation and a tag on it as well as specific keywords related to that. Say for example, if you have something like medical imaging, it's able to annotate what it considers to be risk or no risk and then pass it along to the appropriate healthcare provider. You might want to create an app that is focused on audio and speech components so that people are just able to ask it directly like what were sales in Asia in Q2 like and the app will be able to have a conversation with you about the data. And finally, you can build an internal app for reports and reporting. It cannot be underestimated how much time most companies spend just on generating reports and passing along to either internally or to different stakeholders. You can make an app that automates this process. Everybody will thank you because nobody likes making reports. I think the next category is

hardware related applications. This is a really cool category and I feel like a lot of people don't think about it as much like how much opportunity there is. They're always thinking more just on the software side. But AI is very much able to interact with the physical world. For this type of application, the workflow usually has an emphasis of things being real time and looping. The general workflow is that you collect some type of real-time data usually from some type of physical device like a sensor, a camera, a wearable. The AI would process and analyze that information to detect things like patterns or anomalies or different insights. It will decide if something stands out from that continuous stream of information like trigger an alert or change something. And finally, it would summarize and then report whatever it is that it did. An example of an AI app in this category would be a traffic incident detector. You will have cameras and road sensors feeding in information, continuous information to your AI. The AI would be analyzing that real-time information. And if it detects something like, oh, like there is a stalled vehicle where like a car crash, it would make a decision. Oh, like is this car crash really bad? Is it actually congesting things? And if it does decide that it's a problem, it would then take action. Like it would probably report an alert to an operator and then maybe do things like adjust the traffic lights to try to like mitigate the situation. and then it would finally summarize it saying, "Oh, there's an accident that occurred on like road 7 and then give that to somebody. " Speaking of things related to traffic, you can also make a traffic flow predictor. So, it's able to take historical and live data and be able to forecast different congestions and reroute drivers. Illegal parking detections like having cameras that's able to spark cars that are parked incorrectly and then I guess like be able to find them directly or alert them. crowd management sensors that are detecting the flow of pedestrian traffic and be able to flag different bottlenecks and then like maybe send people who would go, you know, those people who kind of like wave their sticks to redirect people, like get those people to go if there's a lot of congestion. air/noise pollution apps for cities, being able to use sensors and AI to detect where the hot spots are, predict spikes are going to be there, and be able to help like urban planners or like developers figure out where certain buildings should be built and how to like plan the city better. For more general security and camera related apps, you can build something like an object/person recognition application. So, your AI is able to identify packages, pets, vehicles, or known people and be able to flag that. AI privacy filters like it's able to blur out certain things that you don't want to be seen. Sometimes maybe you want to upload a video of something but there's like you know identifying information like people's faces or like license plates and things like that. So you can automatically blur that out. Anomaly detection and camera feeds. The anomaly would be different depending on the different types of scenario but you can get the AI to be able to detect what you consider to be an anomaly in a lot of different types of feeds and CCTV cameras and send it to the popo. There's a lot of applications related to smart home devices, too. Like energy anomaly detection. Maybe you can have smart plugs that's able to, you know, look at your fridge and be like, "Oh, there's a 15% more power usage next month. " And then, I don't know, like miticate that in some fashion. Appliances that would know when it should start itself. Voice activated appliances. Appliances that are able to detect when it should be started and stopped. — Hello, robot. — Yes, I'm here. — Turn on the light. — Okay. a smart fridge inventory that can remind you to buy things or just directly order itself when you know inventory is low on milk. An AI that's able to detect when something is about to fail or some sort of sensor has failed. Like if you have a glucose monitor, if something is going wrong with it, it's able to send an alert. Speaking about wearables and healthcare related things, you can have a wearables aggregator application that's able to take all your different types of wearables like your Aurora link, your Whoop, your Apple Health, whatever. Integrate all of them together and give you insights about your daily activity, how to improve your life. For example, you should stop eating things late at night because it increases your stress response and then you don't sleep very well. A sleep and diet app that's able to figure out what is the right plan for you based upon your current lifestyle. smart accessibility devices like AI enabled hearing aids that's able to increase or decrease certain sounds and filter it so you can hear better. — These systems are a great benefit to the heart of hearing and rapidly increasing around the world. Assisted listening benefits a large and ever growing section of society. Hearing impairment affects one in seven of us, a number that is increasing the population ages. — Smart glasses that has cameras and can like tell you certain things. Kane sensors so you're able to have information about your surroundings. apps that are integrated into your car, being able to tell you when it is that you should get things checked up in a much more direct fashion, not like just an ominous blinking on your car close fleet monitoring like device vehicles or drones. Being able to monitor their different patterns and

then adjust them. The next category is dashboards. AI enhanced dashboards are all about having real time or AI enhanced overviews of specific types of information. It allows users to be able to get an overview of certain systems and then also doubleclick and get more detailed information if they choose to. The emphasis is on how the information is being portrayed for the user. The general workflow of a dashboard starts off with gathering different data sources. And this data source is usually near real-time data sources. This can be through internal information that people are updating. So like CSV files or databases. It can also include external information from different places. Then there's the data processing and analysis which is the cleaning and restructuring of that data followed by the most important part which is the compilation and visualization. So compile everything together showing it to the user. You might also want to help the user out by highlighting like the most important things they should be paying attention to at a glance. And finally distribution. How it is that you want the viewer to be accessing that information. Should it be through an email? Should it just be like a link that you send out to people? A daily report linking back to the dashboard. A classic example of an AI app in the dashboard category is a personal finance dashboard. The app is able to pull the bank statements, different types of transactions from a lot of different places that you're spending your money, process and analyze that information and categorizing the different types of transactions. Then compiling it together and visualizing it like showing you different graphs and charts of how you're spending your money and the different trends over time. It will generate insights for you like, ooh, careful there. You just spent 30% of your monthly budget on chocolate. Probably not a good idea. And finally, distribution. Maybe you want to have like a WhatsApp message from your dashboard saying whenever there's like certain things that are going on, maybe you want an email summary or even like a voice briefing about your spending habits. Another really classic and useful AI app use case is for building a dashboard that can monitor certain trends. For example, if you're a company, you might be interested in your competitor's doings, such as what their marketing stuff is happening, their sales projections, competitor news, industry shifts. In my case, I have an AI news aggregator dashboard that I look to see what are the happenings in the AI world and it's very customized towards things I am specifically interested in. I of course also track to creators in my niche as well to see what kinds of videos and content that they are producing within a company. Building an AI powered KPI internal tracker dashboard is extremely powerful. There is a saying that what doesn't get tracked doesn't get done. You can create a dashboard for customer sentiment looking at all the reviews from different users and then compile them together by product. Anomaly detection dashboard tracking if there's any types of weird spending patterns. Website errors or metric spikes. Cash flow forecasting dashboard. Taking real-time information for accounting and transaction data making sure that we're not overspending or under spending. Pricing intelligence dashboard. Scraping competitor prices and helping you adjust things that are either underpriced or overpriced. Sales pipeline dashboards with CRM integrations. Subscription churns where you're inputting user behavior data. Dashboards for resource allocation. Is the staffing actually correct? Where should we be allocating different projects? Are we really spending money on the right things? Bug/ error prioritization. Which bugs are actually the biggest priority to be addressed? Cyber threat. Maybe there's threatening social media information about your company where there's like attacks that are happening. You want to be monitoring those. And finally, a little passion project that I'm working on, which is an investments dashboard. So, it tracks different assets like real estate or commodities like gold for example, as well as like stocks and bonds. and then it helps me like figure out what I should be investing in and what my trading strategies are. Now, if you're feeling inspired to start building some of these applications, let me introduce you to Bault. It's the first unified platform that takes you from idea to live product in minutes. No code required. Whether you're technical or not, you can go from scratch to launch without worrying about setup or config files. Everything is included right inside Bault Cloud. Hosting, backend, database, file storage, O domains, and even SEO. You can also plug in integrations like Superbase for data and stripe for payments and deploy instantly without ever leaving Vault. It's built to scale too with enterprisegrade performance with partners like Netlefi. You can go from your first user to millions without ever changing platforms. Here are two real examples built using Bolt. Ask in Bio where you can ask creators questions based on their publicly available data and another app that lets you create personalized books and audiobooks for your child. All no code, all running completely on Bolt. Bolt Cloud is currently in private beta and pro users get full access during the beta for free. You should try building your own application based on one of these suggestions right now at bolt. new. It's free and instant. The link is in the description. Thank you so much Bald for sponsoring this portion of the video.

Now back to the video. Next category is chatbot/ass assistance. A lot of apps in this category will have a gent components to it. Not going to go into too much details about building agents right now, but I do have a video which I will link up here that goes into a lot of detail about building agents specifically, but agents are just like a subcategory of AI apps in general. So, all of this is still applicable for chatbots and AI assistants. Um, a really big emphasis here is on the ability to communicate with the user and then also being able to perform actions related to their function. The general workflow of this is that your user will start off with some sort of input, some sort of request. The AI will figure out what the intent of the user is, go and retrieve some information from a database and/or take some type of action and then go back to the user and give it information uh that the user was requesting or informing the user that it has went ahead and performed some type of action. Then finally, logging into behavior somewhere so you know what your chatbot assistant agent has been doing. For example, if you have an accounting assistant, the user can say something like, "Generate an invoice for Lonely Octopus for $5,000. " The accounting assistant would go and classify this request as an invoice creation, pulls the client information for Lonely Octopus, and generate the invoice using some sort of accounting software through their API and then respond to the user saying that the invoice is generated and maybe ask the user, do you want me to go ahead and send this as an email to Lonely Octopus and finally logging that behavior in the internal system so it's trackable. Um, and then if it's sending out an email, also saving the email receipt. You can imagine that you have a legal assistant that could do something similar by generating contracts as well as explaining contracts, explaining clauses, highlighting different risks. An internal policy assistant that's able to answer questions concerning human resources or compliance questions based upon internal documents. A compliance assistant that's able to monitor local laws and alerts you when your business might have to adapt. An invoice assistant that can specifically extract information from invoices and receipts that you might be collecting for different places. And then you can, this is literally like the bane of my existence. So currently in the process of building an AI assistant that can do this. So it's able to take you can take pictures of these receipts and then it would file them automatically for you. Tax filing assistant. Taxes also the bane of my existence. So it can file taxes for you. For small businesses, having a CFO agent could be really helpful because people usually don't like to think so much about that. Um but it's also very important to do things like maintaining and monitoring cash flows and forecasting runway and then giving warnings before payroll. like I have had a lot of experience in messing up these areas. Medical insurance assistant uh especially in the US insurance very confusing having a medical insurance agent that's able to go through that process with you and file claims for you. Benefits navigator if you're working at a company often times you have a lot of benefits that you are not aware of. Medication guide assistant if you have certain prescriptions you can get an assistant to help you explain those prescriptions and fill those prescriptions as well. Health record organizer. You know, if you've ever had the experience of going to different hospitals and they just ask you like the same questions over and over again, you can have an assistant that's able to compile all your health records as a unified health record assistant provider. Customer service agent that's able to handle frequently asked questions and escalate complex task when it's required. A meeting follow-up assistant setting up meetings, just personal assistants. IT help desk assistant with your IT problems. Appointment booking assistants. These are agents that can actually call up different places like restaurants like make reservations or like different appointments with medical appointments and then be able to book certain things for you. Personal shopping assistant, shopping for food, shopping for clothes, shopping for home stuff. Subscription manager assistant. Oh my god, I need this. I have so many different subscriptions and it is very difficult to manage them and I also don't want to pay another subscription to manage my subscriptions. You can just build an agent for that. Warranty and returns assistant. How many times have I just given up and like not returned something because the process of doing that was too complex? Utility negotiator. Sometimes you can negotiate a lot of things like your phone bill, your internet bill, your water bill. There's a lot of things you can actually negotiate and over time it really adds

up. The next category is personalized coaches. So this is also a category that has a lot of agentic components to it but it's different from the general assistance like chatbot assistance because there's a bigger emphasis on the learning aspect and feedback from the user. The general workflow for a coach AI application so your user will be interacting whether like through text, voice, video, action or exercise. The AI would evaluate it like it would score the whatever it is that the user has inputed analyze the user's performance then provide feedback and suggestions for how to improve. It's also important for the coach to be able to provide encouraging reinforcement and then also provide guidance for next steps. And finally, having some sort of progress tracking, be able to track the progression of the user's learning through time. An example of this would be like a relationships coach. The input could be a user being like, I got angry at my partner and I was, you know, not being nice. How do I improve? The AI app will take this information, ask the user questions like, how did they handle this conflict? What exactly was being spoken? What exactly happened? what is it that they spoke about? Evaluate it, then provide feedback for how they can improve, like suggesting better phrasing and being more empathetic. The app will also help reinforce this by suggesting things like roleplaying so that next time they're able to apply these techniques, and progress tracking, seeing if the user has become better in their relationship over time. Some other really cool apps that you can build in this category would be like a drawing coach, being able to upload sketches where just sketching in real time and then having commentary about how to improve. I actually built this one. any type of like sports or physical activity coach like tennis or golf or just like working out being able to upload like video of your golf swing for example and then the coach would be able to analyze your golf swing and then give you like suggestions for how to improve it. Public speaking coach recording a talk or saying what your talk is and then getting real-time feedback and commentary for how to improve. Language learning coach, this one is amazing like having real-time conversations in a specific language and then getting feedback for it. I have also built this one. It's been really great. a career coach that's able to give you mock interview practices and just like any type of like niche specific coach like pottery or like music or like things like this like if you actually hired an expert it would be very expensive to do so. So if you just like build an application to do it for the majority of people they can accomplish like 80 or 90% of their learning through the AI app before having to actually go to a human and pay a lot of money to improve. Next

up is multimodality apps. So we have already seen a lot of like multimodality like text, video, audio and images like a lot of transformations already. But this category of apps I did want to split out by itself because it is such a strong suit of AI. Multimodality apps have a big emphasis on generating and remixing different types of content. The general workflow is that a user will have some sort of input like an idea, a topic or a prompt. The AI app would generate a draft of the content and the user would work with the AI app in order to add different enhancements and different personalizations before finally publishing it and distributing it through maybe like social media or like a newsletter or as an email. An example of an application like this would be like a slide deck generator. The user will have some type of input like make a deck about the history of video games. The app would generate the slide deck the core information surrounding that. Then the user will work with the app in order to enhance it, personalize it, change and tweak different parts of it before finally the user is happy with it and they might export it as a PowerPoint slide, download it as PDF or send it across as an email. This is specifically for creating slide decks, but you can also have a very similar workflow for creating things like newsletters, video content, podcast content, music, images for Instagram, social media posts of any type really. You can also create apps that analyze this type of content like analyzing images and photo trends, analyzing videos, combining these things together, like being able to input video and then having the app comment on the video and give commentary like for a soccer game for example, giving commentary about what's happening in the game. — Okay, I see you're talking about a specific Chrome tab. Is there anything you would like to know about it? — What is happening in this clip right now? Okay, in the video the players are moving around the field and a player in black is on the ground having just been tackled. It looks like the player in red and white has gained possession of the ball. Also, the score is now Arsenal Zero Western — Converter applications like video to text, text to image, image to video, etc., etc. An interactive storytelling app that's able to generate entire stories and have like accompanying text and images and videos and audio and everything. Just playing with these different types of modalities and converting them and combine them together, you can have so many really cool creative ideas and different types of niches as well. Now, time for the

final category, which is automations and macros. This is also a category that is so diverse and I think people don't think about them enough as well. If you think about it, there are like a lot of very repetitive things that we do on a day-to-day basis, both in our personal lives and at work as well. A lot of these things are all opportunities for us to automate and use AI to do it. The general workflow for this type of AI app is to have some sort of trigger that might be like an event or like an input where something basically just happens. Then the AI would take that trigger and maybe like extract that information or take it as a cue in order to perform some type of task. After it runs the specific task, then it will log it and then maybe like communicate it in some fashion to the user. Although for this type of app, the goal is to make it as autonomous as possible. So have it run and do things without direct input from the user. Roughly speaking, there's also two different categories of this type of application. The first type is workflows that are automated on the cloud. For example, you can have a customer feedback classifier. The trigger here would be just having a new customer that arrives through like Zenesk or some type of CRM. The AI app would automatically take that review, extract information about that review, and about the user as well. Take that and label it maybe in terms of urgency, like is it really urgent? Is it something that's very complex? Is it something that doesn't require any action? Then if it's something that maybe is very urgent, requires action, it could do something like go and submit a Jira ticket for customer service to go and fix something based upon the review. And finally, it would have a log of this and then would also report it. The second type of application in this category is usually referred to as macros. And these are applications that are local to your computer. For example, you can have a file organizer. The trigger or input would just be like if you download certain files on your laptop. I don't know about you guys, but mine is like a complete absolute mess. Downloads are everywhere. That would be the trigger for your AI app to take that file, uh, put that into the right folder, and maybe even rename that file so it can be easily be found. Like say if you downloaded an invoice, it would rename that invoice and put that into the invoice folder for your company. Then it would do something like have a record of that and have a report of that. Some other types of AI automations that you can build on the cloud would include something like an automated email repli invoice processing meetinguler sales prospect enricher that's able to take a sales prospect and add additional information like their LinkedIn recruiting pipeline automator that's able to scan and go through uh different potential candidates code review assistant a task generator that's able to convert chat requests into like a Jira or a Trello ticket proposal generator that's able to autodraft proposals. Compliance checker that's checking for specific types of compliance issues and then flagging the things that are problematic. A CRM notetaker that's able to log summaries and then input that into your CRM. Also would be like meeting notetakers as well. A cross tool syncing agent that's able to sync a lot of different types of applications together so you don't need to manually do it yourself. For automation apps that are local, you can have a PDF/doc summarizer that you're able to rightclick onto things and it would instantly summarize things for you. a clipboard assistant, so you're able to copy and paste things and it would expand the text for you. A voice command agent, so you're able to have the ability of communicating with your computer directly through voice, like telling it to run certain scripts. A spreadsheet macro agent that's able to input the common formulas that you usually use on your spreadsheet. Screenshot analyzer. If you take a lot of screenshots, it might automatically be able to just take the screenshot and annotate it and then put it in the right place. Local search assistant, maybe you have a lot of files and documents and things like that and you want to search across your local computer. So, you can create an agent that's able to do that for you. A photo tagger. Have a lot of photos, want to attack the photos. Presentation helper. Maybe you're creating a presentation. You wanted to stay on your computer for security reasons. A security assistant that's able to scan your local files before you upload it onto the internet to make sure you're not revealing any of your secrets. So, there you have it. That was a lot. There are so many different types of AI applications that you can start building. I really hope that by watching this video, you have some inspiration now and there's certain things that you haven't thought about and that you're really excited to start building. Let me know in the comments what it is that you want to build out and I will see you guys in the next video or live stream.