MakeToday

MakeToday gives you one personalized pick for what to do based on your mood, time, budget, and location—so you spend less time deciding and more time doing.

Comment

Inspiration

I created MakeToday because I feel like people often get stuck between wanting to do something and figuring out what to do. Part of the problem lies in the abundance of options, which can overwhelm us. Narrowing down the number of options can help remove some of the pressure of deciding. Other times, it can help give us a push if we feel like we are doing something that we did not choose entirely by ourselves. It could even be as simple as wanting a new idea or a new place to see. I wanted to create a simple app that tailors its recommendations to the user based on various factors surrounding their situation.

What it does

MakeToday gives users personalized activities based on their current situation. Instead of give the user and endless list, it gives three personalized recommendations.

I chose three because I still wanted users to have control. Giving only one answer could feel too restrictive, but presenting a long list would turn deciding into another task. Three choices provide flexibility while keeping the experience focused. Users can tailor their recommendations using details such as:

  • Mood and energy
  • Available time
  • Budget
  • Location
  • Preferred travel time
  • Transportation methods
  • Whether they are alone or with one or two other people
  • When they want to begin

It is already easy to search "things to do" on Google, but those searches often produce another long list to sort through with little information. MakeToday considers the user’s actual situation to remove unsuitable choices before presenting anything. This makes the results more personalized and increases the chance of the user choosing said choice.

Each recommendation provides practical, easy-to-scan information, including the location, suggested start or departure time, activity duration, estimated cost, travel details, weather considerations, opening information, and links to the official website and Google Maps if available. Users can swipe through their three options, and select "I'm doing this" to lock in their choice for the day or generate more options up to three times total for the day. After selecting their choice the user will be prompted whether they have done it or not.

MakeToday also includes My Days, an optional journal where users can save what they did, add a picture, and write about the experience. Looking back at even small activities can make a day feel more memorable and give someone a sense that they accomplished something. It can also provide a foundation for optional sharing, allowing people to share activities with others while keeping participation under their control.

Design

Weather can determine whether an activity is a good choice, so I wanted it to influence both the recommendations and the appearance of the app. The prototype includes different weather scenarios to demonstrate how this could work with live weather data in a future version.

I wanted the interface to remain simple so using it would not feel like another chore or overwhelming. The weather scene continues behind the choices, to create a smooth transition and keeping the experience visually connected instead of making every section feel like a separate page.

Although the question “What should I do today?” applies to a broad audience, I think MakeToday could be especially useful for people who struggle more with decision-making, anxiety, or finding the push to get out and do something. It is not intended to focus on those who may suffer from anxiety and depression, but I think it may benefit those people the most.

How I built it

I began this project without coding knowledge. I directed the concept, appearance, features, interactions, recommendation behavior, and repeated revisions. Codex handled the technical implementation by creating and modifying the code, building the recommendation logic, writing and running tests, identifying errors, and helping prepare the application for deployment.

My role was to describe the intended experience, review every version, test it as a user, identify what did not feel right, and decide what should change. I frequently reviewed the running interface and gave Codex specific feedback about animations, transitions, typography, recommendation cards, navigation, content, and mobile usability.

GPT-5.6 and Codex allowed me to turn those product decisions into a working application. Codex accelerated the entire development workflow, but the project still required ongoing human judgment and direction rather than a single prompt.

The Build Week version uses a curated set of places and at-home activities with deterministic local recommendation logic. It evaluates eligibility and scores candidates using the user’s choices, then selects three varied recommendations. Weather, travel estimates, opening information, and place data are simulated or curated for the prototype rather than retrieved from paid live services. This keeps the demonstration reliable and makes its current limitations transparent.

The application was built as a mobile-first web experience using Next.js, TypeScript, and Tailwind CSS. User choices, active plans, preferences, usage information, and My Days content are stored locally in the browser.

Challenges

The biggest challenge was that nearly every part of building and publishing an application was new to me. Because I began without coding knowledge, I could not manually open a file and confidently change the implementation myself.

Instead, I had to learn how to communicate what I wanted clearly, test each version, identify what did not feel right, and give Codex increasingly specific instructions. I also encountered unfamiliar concepts such as repositories, deployment, browser storage, testing, and responsive design.

I would not describe the process as frustrating because I approached it as a new experience. However, it required me to understand how the different parts fit together while the application was being built.

Accomplishments that I’m proud of

I am most proud of both the concept and the final appearance of MakeToday. The product has broad everyday usefulness while potentially providing an even greater benefit to people who find deciding or getting started more difficult.

I am also proud that I turned an idea into a working, published prototype despite beginning with no coding experience. The finished application includes a complete decision flow, recommendation logic, responsive results, commitment tracking, profiles and preferences, usage limits, and the My Days journal.

What I learned

I learned that creating an application with AI is not simply describing an idea once and receiving a finished product. Seeing and testing each version helped me notice features, interactions, and design details that I would not have considered while the idea existed only in my head.

Reviewing the application gave me something concrete to react to, and each revision made my vision clearer. I learned that my role was not only to provide the initial idea, but also to test the result, question decisions, explain what felt wrong, and guide the product through many smaller improvements.

What’s next for MakeToday

My next priority is connecting MakeToday to live services so it can work beyond its curated prototype data.

Place photographs are especially important because users should be able to see where they may be going and get a sense of the experience before committing. I also want to add AI-assisted ranking so the app can understand more nuanced input and select the strongest recommendations from verified nearby places.

Future development could include:

  • Live nearby-place information and photographs
  • Current and hourly weather
  • Travel times for different transportation methods
  • AI-assisted ranking of verified candidates
  • User accounts and cross-device profiles
  • Improved personalization based on completed activities
  • Optional sharing through My Days

The AI would rank verified information rather than inventing places, hours, prices, weather, or travel estimates. My long-term goal is to preserve the simplicity of the prototype while making every recommendation more accurate, personal, and actionable.

Built With

Share this project:

Updates