Inspiration

Founders shouldn't have to keep fifty tabs open just to figure out which accelerator to apply to. The info is all out there it's just scattered, inconsistent, and exhausting to compare. We wanted an agent to do that part.

What it does

Fun Dip onboards you with a quick chat, builds your company profile, scouts the web for programs that fit your stage and goals, and applies on your behalf when you pick one. If the application needs context the profile doesn't have, the agent asks once and remembers it.

How we built it

React on the front, WunderGraph routing the API, RocketRide running the AI pipelines (profile, programs, submissions), and Ghost as the agent-friendly Postgres database. TinyFish handles the browser agents that actually visit program sites and fill out forms.

Challenges we ran into

Application forms are wildly inconsistent, every program site is its own little kingdom. Getting the agent to know when to ask the user versus when to just figure it out took a lot of tuning. Schema design was tricky too, since "program" means something slightly different for accelerators, VCs, and partnerships.

Accomplishments that we're proud of

The end-to-end flow actually works. Onboard, discover, apply, all in one place. And it doesn't look like every other AI app; the workspace feels calm.

What we learned

Agentic UX is a layout problem as much as a model problem. Where the chat lives on the screen changes how people use it.

What's next for Fun Dip

Deadline reminders, smarter fit ranking based on what actually lands interviews, and expanding beyond programs into grants and partnerships.

Built With

Share this project:

Updates