Skip to content
docs
UseBuildChangelog
Open Lumin

Use cases

What you can build on Lumin

98 use cases across 19 verticals, each grounded in tools that actually exist on the server. Every tool name below is real and links to its reference page. Roughly a third of the surface is not orthodox KP (Parashari, Jaimini, Tajik, KP-extended), and presenting one of those as a KP finding is a methodology error that reads as thoroughness, so each vertical page labels it.

Start here: what you can ask the user for decides what you can build

This is the most useful fact on this page. Tools divide by input class, and the class decides the product shape, the consent burden, and the conversion friction.

Input classToolsWhat it needsConsent friction
Natal172Birth date, time and placeHigh. Birth data is among the most identifying tuples a product can hold, and a large share of users do not know their birth time.
Place and/or date only11A location and a date. No person at allNone. Cacheable per city per day, embeddable on a public page.
Horary7A number 1 to 249 and the moment of askingNone. Answers real life questions with no birth data.
Two-person9Two chartsHigh, doubled
Mundane2A country, company or location chartNone, no private individual
Reference3NothingNone. Free discovery calls

Two corrections

Both verified against the engine. Assuming either from the family name alone will misroute a product decision.

  • The electional family is not place-only. 14 of its 15 tools take the native's birth data, because KP election consults the running dasha lord of that person's chart. Only get_election_catalog is person-free.
  • The horary family takes birth-data-shaped field names but no birth. birth_datetime, latitude and longitude carry the moment and place of judgment, not a birth.

The seven verticals

Every one of the 98 use cases in the catalog appears on exactly one of these pages, mapped from the source catalog's 19 lettered sections.

PageCatalog sectionsUse casesCovers
Matrimonial and datingA8Match scores, dosha filters, wedding dates both families accept.
Commerce and personalizationB, O12Onboarding profiles, gift finders, send-time optimisation, a personalization API.
Wellness and healthC, D13Prakriti quizzes, constitutional risk profiles, recovery windows, triage.
Career and educationE, F12Vocational fit, promotion prep, exam windows, study-abroad counselling.
Scheduling and electionalJ, P, Q12Venue dates, meeting and filing slots, panchang widgets, regional almanacs.
Place-based and operationsI, K9Departure timing, sowing and harvest windows, fleet overlays, zero PII.
Research and dataG, H, L, M, N, R, S32Fintech, real estate, sports, legal, media, research and consumer apps.

Before you ship any of these

Read responsible use first. It is the compliance page: disclaimer requirements by vertical, the three advisor-only tools that must never reach a consumer, and why birth data is PII.

Read pagination.totalItems and pageNote on paged tools, and increment
page until you have what you need. A page is a unit of thinking, not a
payload optimisation.

Calls are topped up in any amount, starting at 1 USD for 400 calls.
A quick lookup is a handful of calls and a full reading is 25 to 40,
so more calls means more readings and deeper ones. The eight run_*
composites chain several engine calls into one, which covers the same
ground for fewer calls when your UI does not need the intermediate
results.

Where to next