Momo: Become Local in 30 Seconds
The hardest part of travel is not always language.
It is culture.
You might be standing in front of a menu, unsure what to order.
You might walk into a temple, worried that one wrong move could offend someone.
You might take a photo of a beautiful building, only to notice people around you quietly bowing, and you have no idea why.
Modern AI can translate menus, recognize landmarks, and convert text between languages.
But it rarely answers the questions that travelers actually feel in the moment:
- Why do locals do this?
- Is there any etiquette I should know?
- What should I avoid?
- What is the story behind what I am seeing?
Those answers are often what turn a trip from sightseeing into understanding.
Translation explains words. Culture explains meaning.
Take one famous Sichuan dish: Fuqi Feipian.
A translation app might render it as:
“Husband and Wife Lung Slices.”
Technically, the translation is not the point. The important part is the story: why the dish has that name, what it means in Chengdu food culture, and why generations of people still cherish it.
That was the moment behind Momo.
We did not want to build another translator.
We wanted to build a cultural companion.
What we built
Momo — Become Local in 30 Seconds turns a travel photo into a playful cultural discovery.
Snap or upload a photo of a dish, temple, street sign, ritual, landmark, or everyday gesture. Momo uses a small multimodal model to decode what you are seeing into four layers:
What tourists see
A clear description of the visible object or scene.What locals know
The hidden rule, local context, or cultural detail most visitors miss.What not to do
Practical etiquette, taboos, or mistakes to avoid.Why it matters
The deeper history, emotion, or social meaning behind the moment.
Why it belongs in Thousand Token Wood
Momo is not just an identification tool.
It is a small cultural adventure: every photo becomes a discovery, every discovery can become a memory, and every memory contributes to a growing cultural passport.
We designed the experience around the feeling of travel:
- discovery
- collection
- local stories
- respectful curiosity
- a sense of progress
The interface is a custom Gradio app with a playful travel UI: World, Passport, Compass, Journey, and Traveler.
Small models, real product
Momo is built for the Build Small Hackathon using small open models.
Core model:
openbmb/MiniCPM-V-4.6for image understanding and cultural decoding
Experimental narration backend:
openbmb/VoxCPM2
Both models are far below the 32B parameter limit.
The app is hosted as a Gradio Space inside the official Build Small organization.
What we learned
The most interesting travel question is not always:
“What is this?”
It is often:
“What does this mean to the people who live here?”
Many people travel to collect photos.
We want to help them collect stories.
Many AI tools translate language.
We want Momo to translate culture.
Thirty seconds may not be enough to learn a new language.
But it might be enough to see the world through local eyes.
Links
- Space: https://huggingface.co/spaces/build-small-hackathon/local-in-30s
- Demo video: https://youtu.be/ixmp0grzGtU
- Social media post：https://x.com/i/status/2066069050481066303
Built for Track 2 — Thousand Token Wood.
Prize targets: Best MiniCPM Build, Off Brand / Custom UI, Field Notes, and Best Demo.