vedicproof
vedic proof · computation

Kundli Matching API

What is a Kundli matching API?

A Kundli matching API compares two birth charts and scores their compatibility. The standard method is Ashta Kuta, or Guna Milan: eight kootas worth thirty-six points in total, derived mainly from each person's Moon nakshatra and rasi. The API returns the total and each koota separately.

Request

Tools are called over MCP, so an assistant connected to https://vedicproof.com/mcp invokes check_compatibility itself. The arguments it sends are:

tool · check_compatibility
{
  "boy": {
    "date_of_birth": "1992-06-11",
    "time_of_birth": "14:05",
    "city": "Pune",
    "country": "India"
  },
  "girl": {
    "date_of_birth": "1994-01-19",
    "time_of_birth": "05:22",
    "city": "Kochi",
    "country": "India"
  }
}

Response

Every response carries a plain-language summary string plus structured data, and states the engine and ayanamsa it used. Abridged here for length.

result · json
{
  "summary": "Guna Milan 23/36. Moon nakshatras Chitra and Shatabhisha. Nadi koota clears; Bhakoot does not.",
  "total": 23,
  "maximum": 36,
  "kootas": [
    { "name": "Varna", "score": 1, "maximum": 1 },
    { "name": "Vashya", "score": 0, "maximum": 2 },
    { "name": "Nadi", "score": 8, "maximum": 8 }
  ]
}

How is this calculated and verified?

These are the pinned parameters that affect this computation. The full disclosure, and the golden-chart corpus, live on /proof.

AyanamsaLahiri (Chitrapaksha), mean
Single supported system today, computed as a polynomial in Julian centuries from J2000 rather than read from a table. True-Chitra, KP-Newcomb and Raman are not offered; that is a limitation, not a default we are hiding.
Rounding and boundariesAssigned on unrounded longitude; displayed to the arc-minute
Rasi, nakshatra and pada are assigned from the full-precision longitude before any rounding, so a graha at 29.9999° stays in its own rasi. Degrees are formatted d° mm' ss" for display only.
Ephemeris source and versionBuilt-in analytic series - not Swiss Ephemeris
Sun from a truncated VSOP-style series, Moon from truncated ELP2000 terms, and Mercury through Saturn from JPL Keplerian elements. Accuracy is arc-minutes, not arc-seconds: a few arc-minutes for the Sun and Moon, up to roughly a tenth of a degree for the outer grahas. That is fine for rasi, nakshatra, pada, dasha and koota output and is not fine for arc-second work. Every API response carries the engine note stating this.

Common questions

Is a score out of 36 enough to decide a match?
It is one traditional filter, usually read alongside Manglik status and the seventh house. The API returns the breakdown precisely so a product can show which kootas failed.
What score is normally considered acceptable?
Practitioners commonly treat 18 and above as workable, with Nadi and Bhakoot weighted heavily regardless of the total. The API reports; it does not judge.
Are birth times required?
For Ashta Kuta the Moon matters most, so an approximate time still gives a usable score. Without any time the API assumes noon and flags the assumption.

Connect this to Claude or ChatGPT with one URL: https://vedicproof.com/mcp. First 14 days unlimited and free, then $1.99/month.

claude setup →chatgpt setup →pricing →

Related computations

all computations in docs →