Inspiration

We're all constantly making tiny predictions — "this meeting is going to run over," "that stock's going to pop," "my sister's totally going to get engaged this year" — but they vanish the moment we think them. We wanted to capture that instinct and turn it into something fun and communal: a place where people commit to a guess about the future, in whatever form feels natural (a doodle, a scribbled note, a voice memo), and then get to see, together, who called it and who didn't.

What it does

Concert of Futures lets you record a prediction — by drawing it, writing it by hand, or speaking it into a voice note — and attach it to a specific future date and time. Once submitted, your prediction is locked and added to a public timeline alongside everyone else's. As each prediction's target time passes, the community (or the original predictor) can mark it as right, wrong, or "eh, sort of." You can browse past predictions by date, person, or topic, watch a leaderboard of the most accurate predictors, and relive old guesses to see just how wildly right or wrong we all were.

How we built it

We built a lightweight web app with a canvas-based drawing input for sketches, a handwriting capture mode, and audio recording for voice notes. Each prediction is timestamped and stored with its target "reveal" date. We used [insert your actual stack — e.g., React frontend, Node/Express backend, a database like Firebase or PostgreSQL, and speech-to-text for voice note transcription] to tie it all together. A simple scheduler flags predictions as "ready to judge" once their target time arrives, surfacing them for resolution.

Challenges we ran into

Handling three very different input types — drawings, handwriting, and audio — in one unified prediction format was trickier than expected, since each needed its own capture UI but had to feel consistent in the final feed. We also had to think carefully about how to fairly resolve predictions (who decides right vs. wrong?) without making the app feel like extra work, and we spent a lot of time getting the timeline/feed UI to feel alive rather than like a static archive.

Accomplishments that we're proud of

We're proud of getting all three input methods working smoothly in the time we had, and of the "reveal" mechanic that makes checking old predictions feel like a fun payoff rather than a chore. We also liked how quickly a browsable, communal feed of predictions started to feel genuinely entertaining once we had even a small amount of test data in it.

What we learned

We learned a lot about handling multi-modal input (canvas drawings, audio blobs, and text) in a single consistent data model, and about the UX challenges of designing for delayed gratification — since the "fun" of a prediction app only shows up once time has actually passed. We also learned how important lightweight, low-friction input is: if predicting something takes more than a few seconds, people won't do it.

What's next for Predict the Future

Next, we want to add social features like following specific predictors, prediction "streaks," and topic-based leaderboards (e.g., tech predictions vs. personal life predictions). We'd also love to add AI-assisted resolution — using speech/handwriting recognition plus real-world data feeds to auto-suggest whether a prediction came true — and push notifications that remind you right when one of your old predictions is about to be revealed.

Built With

• react
• reve