BrainDump

Inspiration

As college students, we sometimes have so many assignments, errands, and deadlines on our minds that making a to-do list feels like another task. It’s easier to say everything out loud, even if it comes out a little messy. We built BrainDump to turn those thoughts into an actual list.

What it does

BrainDump turns text dictated through Voiskey into an editable task list. Using AI, BrainDump identifies individual tasks, interprets deadlines and priorities, and suggests manageable first steps to help users get started.

Users can edit tasks, check them off, filter by deadline or priority, and export their plan. They can also dictate simple corrections, like “Move laundry to tomorrow.” Tasks are saved in the browser so users can return to them later.

How we built it

I built BrainDump using HTML, CSS, JavaScript, and a Hugging Face LLM, with AI coding assistance to help develop the prototype during the hackathon.

Voiskey handles speech-to-text, allowing users to speak their thoughts directly into the website's text box. BrainDump uses a Hugging Face LLM to interpret the user's input, identify individual tasks, recognize deadlines, and assign priorities. The AI also suggests actionable first steps for each task, making it easier for users to get started.

Challenges we ran into

Another challenge was making sure the AI could accurately interpret messy, unstructured thoughts. For example, "buy bread and butter" should remain one task rather than being split into two. We also had to consider how the model would interpret deadlines and corrections. Since AI-generated results aren't always perfect, we made tasks editable so users could easily make adjustments.

Accomplishments that we're proud of

We're proud of getting a working project online within a short hackathon window. Beyond creating tasks, BrainDump supports editing, completion tracking, undo, and simple dictated corrections.

We also wanted the interface to feel calm and approachable. The goal was to make a busy day feel a little more manageable.

What we learned

We learned how browser storage can keep a small application useful without a backend, and how much attention seemingly simple features like date recognition need.

Working with AI coding tools also showed why testing matters. Checking the behavior and understanding the code's limits are vital parts to completing the project.

What's next for BrainDump

We'd like to improve our existing AI integration so BrainDump can understand more complicated thoughts, handle conversational corrections, and ask follow-up questions when something is unclear. Calendar integration, reminders, and optional cross-device syncing would also make the app more useful.

If Voiskey provides a supported API or SDK, we'd love to explore a more direct integration.

Built With

• ai
• css
• html
• huggingface
• javascript
• llm