Kundli matching in four layers, not one percentage
Guna milan reads one variable — the two Moons. That is a useful filter and a poor verdict. This runs it as the gate it classically is, then does the three things a "82% compatible" score hides.
Free, no sign-up. Every calculation runs in your browser — your birth date, time and place are never sent to a server.
The four layers
| Layer | Question it answers |
|---|---|
| 1 · Ashtakoot guna milan | Are the two temperaments rhythmically compatible? Eight kootas out of 36 — Varna, Vashya, Tara, Yoni, Graha Maitri, Gana, Bhakoot, Nadi — with the classical cancellation rules applied. |
| 2 · Dosha analysis | Does either chart carry a structural affliction to marriage, independent of the other person? Kuja (Mangal) dosha graded across three reference points, with sign exemptions and mutual cancellation. |
| 3 · Marriage promise | What does each chart say about partnership before the other person exists? The 7th house and its lord, Venus, the patikaraka, the navamsa, the darakaraka and the Upapada Lagna. |
| 4 · Synastry overlay | What happens when these two specific charts are laid over each other? Directional, orb-weighted contacts — the layer almost universally missing from matching apps. |
Why there is no overall percentage
Because the useful information is where the layers disagree, and averaging destroys exactly that. A pair can score 29.5 out of 36 on temperament and still have an adverse chart overlay; that is a specific, explicable, actionable result. Flattening it into one number throws away the only thing the analysis produced.
So the report names the conflicts instead: temperament aligning while structure does not, a non-mutual overlay, a chart whose own promise is weak regardless of the partner, doshas that cancel each other.
Compatibility is not symmetric
A onto B is not B onto A. One person's Moon can land on the other's ascendant while the reciprocal Moon lands in their sixth house — one partner feeling met by the relationship while the other does not, with neither doing anything wrong. No symmetric score can represent that, so both directions are computed and shown separately.
Guna milan is a gate
Classically the 36-point count is a preliminary filter: 18 or above and you proceed to the real analysis. It is not a grade. A 32 and a 21 route into the deeper layers identically, and the deeper layers carry more predictive weight — which is why a high koota score with a combust 7th lord is a worse position than a middling score with a clean one.
Cancellations are not optional
Omitting the parihara rules is the largest single source of false alarms in matching tools. Gana dosha is void when the Moon-sign lords are mutual friends. Bhakoot is void when they share a lord. Nadi has four separate routes out. Kuja dosha has sign-specific exemptions per house and cancels mutually when both charts carry it. All of them are applied here, and the report says which one fired.
Common questions
- What is a good kundli matching score?
- 18 out of 36 is the classical threshold to proceed — but treat it as a gate rather than a grade. It measures temperament between two Moon positions and nothing else: not either chart's own marriage indications, not afflictions, and not what happens when the two charts are overlaid.
- Is Mangal dosha a deal-breaker?
- Less often than it is presented as. It is graded here across three reference points, and it has real cancellations: Mars in its own sign or exalted, a Jupiter or Saturn aspect, sign-specific exemptions per house, and mutual cancellation when both charts carry it.
- Why does this not give me a compatibility percentage?
- Because a single figure cannot represent four layers that routinely disagree, and the disagreements are the informative part. A pair with strong temperament and an adverse overlay would read as a high percentage and be actively misleading.
- Do you need both birth times?
- For the koota score, only the two Moon positions, so an approximate time often suffices. For layers 2 to 4 — doshas, marriage promise and the overlay — both exact times are needed, because all three depend on the ascendant.