This is a submission for the OpenClaw Challenge.
What I Built
This idea started after two young tourists asked me whether the subway at this station was going uptown or downtown. Although there were signs, they did not speak English well and thus were unsure which way the subway was going. FindMySubway is a concept for a subway companion tool designed to help people confidently choose the right train and direction, while navigating the NYC subway system. Other apps like Google Maps and the MTA app are great at showing routes, but they lack in helping the rider determine which platform to stand on, and which subway makes the most sense to take right now. This idea focuses on clarity over complexity. Instead of listing routes, it helps people understand what subway to take by showing where they are, clearly explaining direction, and providing results in a simple and friendly language. Think of it like your bff tour-guide in a pocket!
How I Used OpenClaw
I used OpenClaw as a reasoning and explanation layer between transit data and the rider.
- Interpret natural language requests such as “I want to get to SoHo”
- Determine the correct subway direction based on the rider’s current location.
- Compare available trains and explain which option is best in the moment.
- Translate transit jargon into clear, human‑friendly instructions. Instead of presenting a long list of routes, OpenClaw generates context-aware guidance like: “You need to go downtown. The D train arriving in 1 minute is your best option. If you miss it, the B train in 6 minutes will also work.” This approach prioritizes confidence and understanding, especially for tourists, new residents, or tired daily commuters who may struggle with platform layouts and directionality.
Demo
The demo walkthrough shows how a user enters a destination in natural language and receives clear, step‑by‑step guidance before boarding a train.
It highlights:
- Natural language destination input.
- Clear identification of train direction (uptown vs downtown).
- Plain English explanations for why a specific train is recommended.
The demo focuses on the user decision moment rather than a full production app, emphasizing experience and clarity over polish.
Demo:
Initial prompt:
"I am at Prince Street in Soho, and I need to get to Union Square."
FindMySubway response:
"Best route: Take the Uptown R or W from Prince St to 14 St-Union Sq.
Enter here: Use the Uptown/Queens-bound entrance at Prince St; if entrances are split, take the Uptown side.
Take: Uptown R or W
Transfer: No transfer
Get off: 14 St-Union Sq
Backup: If you are closer to the 6 at Spring St or Bleecker, that can also work, but Prince St on the R/W is usually the simplest."
This initial response sounded a bit too robotic. After some updates, it sounds a bit better. I think it could still be improved, but we're getting there!
Prompt: "I am at Prince Street in Soho, and I need to get to Union Square."
FindMySubway response: "Take the Uptown R or W from Prince St to Union Square.
Use the Uptown/Queens-bound entrance at Prince St — if the entrances are split, make sure you take the Uptown side.
No transfer — get off at 14 St-Union Sq.
If you are actually closer to Spring St or Bleecker, the 6 can work too, but R/W from Prince St is usually the easiest move."
What I Learned
One of the biggest takeaways from this project is how difficult subway navigation becomes under stress, even when accurate data is available.
Working with OpenClaw highlighted the importance of good prompts and contextual reasoning, especially when transforming complex systems into guidance that feels genuinely helpful to people in real-world situations.
ClawCon Michigan
n/a
Top comments (0)