Loolulala by MUTO · In development

A better place
to meet.
Or part ways.

You are already sharing the ride. Where should you meet or part ways? Loolulala is being built to compare the driver's extra effort with the passenger's walk or transit journey.

Try the comparison

룰루랄라 — 운전자도, 동승자도 편한 승하차 지점.

Pickup + drop-off codeMobile prototypePublic concept demo
Two journeys. One shared point.Concept
A shared drop-off point along a driver's route An illustrative, non-geographic diagram. A green driving route meets a dotted passenger path at a shared point. No live map data is shown. StartDriver destinationPassenger destination A shared point
Driver’s routePassenger’s onward journey

Illustration only · not a real route

The idea

Share the ride.
Balance the effort.

01

Bring both journeys

Start with the driver’s route and the passenger’s destination — or their starting point for a pickup.

02

Compare the trade-offs

Consider extra driving alongside walking and transit. The closest point on a map may not make the easiest journey.

03

Choose together

Review the options and agree on a point. Stopping suitability and pedestrian access need separate confirmation.

Interactive concept

What matters more
on this journey?

Try a comparison using fictional sample times. This is an explanation of the product idea, not a live route recommendation or a Claude response.

Journey

Emphasis

Point A

Sample option

Keep the driver close to their route

Extra driving
1 min
Onward walk
12 min

Point B

Best sample fit

Share the extra travel effort

Extra driving
4 min
Onward walk
6 min

Point C

Sample option

A shorter walk for the passenger

Extra driving
9 min
Onward walk
3 min

Point B best fits equal emphasis in this sample: 4 minutes of extra driving and a 6-minute onward walk.

The example uses a simple weighted time score and makes no map or Claude API calls. Real recommendations will also need route data, arrival timing, access checks and your constraints.

What is built

A real foundation.
A clear next step.

Loolulala builds on our Waypoint prototype: mobile screens, pickup and drop-off calculation handlers, and route-service integrations. We are developing the product before a public route-planning launch.

See the development plan ↓
Code implemented

Two directions, shared travel effort

Pickup and drop-off flows include driver-focused, balanced and passenger-focused candidate categories. Flutter mobile screens and backend calculation handlers exist.

Integration code implemented

Driving, walking and transit

Adapters connect TMAP driving and walking, ODSAY transit and TAGO bus-stop data. The live services and a complete end-to-end journey still need validation.

Local checks passed

109 tests across 8 suites

The current prototype passed its local automated tests and TypeScript check on 9 October 2026. These checks cover code behavior; they do not establish real-route accuracy or field performance.

Next validation

Measure the actual extra journey

Validate the time difference between the original and candidate routes, distinct alternatives, arrival timing and pedestrian access. Stopping suitability must be checked separately.

Why Claude

People describe a situation.
Routes need clear constraints.

“Somewhere convenient” means different things to each person. Our planned Claude layer turns that conversation into usable preferences, asks for missing details, and explains the trade-offs in computed route options.

Illustrative request

“운전은 5분까지 더 해도 돼. 친구는 짐이 있어서 7분 이상 걷기 어렵고, 지하철로 갈아타도 괜찮아.”

“I can drive up to 5 extra minutes. My friend has luggage and can walk up to 7 minutes; taking the subway is okay.”

Proposed structured interpretation
Extra driving limit
5 minutes
Walking limit
7 minutes
Transit
Allowed
Clarification needed
Exact destinations and departure time

This is a prepared example of the intended behavior, not a live model response.

01 / Understand

Extract and clarify

Use Claude to interpret Korean-language preferences and return explicit constraints. Ask a focused question when a destination, time or requirement is unclear.

02 / Calculate

Ground the options in route data

Routing services provide candidate routes and travel times. Code calculates extra effort, applies limits and reports when no candidate meets the requirements.

03 / Explain

Make the trade-off readable

Use Claude to explain each computed option in everyday language, including missing data and why an option does or does not fit. Never invent travel times or stopping rules.

Integration status: Claude is planned and has not yet been connected to the live product. We will evaluate constraint accuracy, useful clarification, explanation fidelity, latency and cost before a pilot.

Development plan

From prototype
to a tested journey.

The next release should earn trust with a small, complete workflow.

  1. Foundation exists

    Pickup and drop-off prototype

    Mobile screens, route handlers, service adapters and this interactive concept are the starting point.

  2. Next

    Validate route comparisons

    Check real detours and onward journeys, distinguish the candidate options, and handle unavailable data and infeasible constraints.

  3. Then

    Evaluate Claude, then run a pilot

    Test ambiguous and conflicting Korean requests against expected constraints. Compare model explanations with route results and collect feedback from consenting pilot participants.

The team behind Loolulala

Loolulala · 룰루랄라

A product in development by MUTO (주식회사 카페무토),
founded in February 2026 and led by Lee Hyun Ki.

Built on the Waypoint prototype, now being brought into MUTO's product development.

About MUTO ↗

Product and pilot enquiries

hkyisland@loolulala.com