Calculation you can check.
Drik Gaṇita is the observational school of Vedic astronomy — the tradition that insists a computed position must agree with the sky you can look at. It is where the name comes from, and it is the standard the whole engine is built to.
दृक् · dṛk
Sight. The eye. What is actually seen, as against what is merely inferred.
गणित · gaṇita
Calculation. Reckoning. The mathematical half of Jyotiṣa.
दृक्गणित · dṛk-gaṇita
The observational system: positions computed to match the observed sky, not inherited tables.
01
Pipeline
From birth data to a defensible statement
Six stages. Each one narrows what the next is allowed to say.
01
Ephemeris
Positions
Birth data goes in; the Swiss Ephemeris gives geocentric positions, and the selected ayanamsa converts them to sidereal. From there: Lagna, whole-sign houses, nakshatra and pada, dignity, retrogression, and every divisional chart from D2 to D60. D60 in particular is computed at sub-degree precision rather than rounded — it is the varga most software quietly approximates.
02
Timing
Dashas and transits
The Moon’s nakshatra at birth fixes the start of the 120-year Vimshottari cycle, and the full mahadasha, antardasha and pratyantardasha sequence is generated from it — never inferred, never rounded to a year. Narayana and Shoola dashas, current transits, Sade Sati phase, monthly transit tables and panchang all come off the same clock.
03
Rules
Yogas and doshas
A deterministic rule engine reads the computed chart and applies classical combinations directly: 112 rules across fifteen categories, including the Pancha Mahapurusha set, Raja and Viparita Raja yogas, Neecha Bhanga, Daridra and Kemadruma. Each detection reports the planets and houses that triggered it and the classical reference behind the rule. Nothing downstream is free to invent a yoga the rules did not find.
04
Corpus
Classical sources
Classical texts are chunked verse by verse, tagged by subject, and embedded for retrieval — so a question about Mars in the seventh house retrieves the verses that actually address it, by identifier, not by resemblance.
05
Synthesis
Interpretation, then verification
Written interpretation is composed against the fixed chart facts, the detected yogas and the retrieved verses — and then checked. Every position it states is compared back to the computed chart, and every citation is resolved against the corpus. A mismatch regenerates; a persistent mismatch goes to human review rather than out the door.
06
Regression
The golden dataset
A reference set of documented charts — public birth data, cross-checked against established Jyotish software and signed off by practitioners, each with real dated life events and their active dasha lords — runs as an automated suite. A calculation regression blocks the deploy. That is the whole quality gate, and it is not optional.
02
Constraints
The rules we hold ourselves to
Deterministic by construction
Positions come from the Swiss Ephemeris and a pure calculation layer with no inference anywhere in the path. The same birth data returns byte-identical output today, next year, and a decade from now.
Yogas are detected, not guessed
A rule engine reads the chart and applies classical combinations directly. Nothing in the interpretation path is free to invent a yoga that the rules did not find.
Every claim carries a citation
Interpretation is retrieved against a chunked corpus of classical texts and each statement resolves to a specific shloka. If no supporting verse retrieves, the claim is not made.
Verified against a golden dataset
A reference set of documented charts, cross-checked against established Jyotish software and signed off by practitioners, runs as a regression suite. A calculation regression blocks the deploy.
What the engine guarantees is arithmetic, not prophecy. Positions, dashas, vargas and yoga detection are deterministic and testable, and we defend them. Interpretation of a five-thousand-year-old tradition is a matter of judgment, and qualified practitioners differ — so the engine cites what it is drawing on and leaves the reading legible rather than pretending to settle it.
03
Questions
Asked and answered
01What does "Drik" mean?
Drik (दृक्) means sight, or seeing. In Vedic astronomy, Drik Ganita is the observational system of calculation — positions worked out from where the sky actually is, rather than read off inherited tables. That distinction is the whole thesis of this engine.
02Which ayanamsa do you use?
Lahiri is the default and the primary, matching the Government of India standard. True Chitra, Raman, KP, Pushya-Paksha, Surya Siddhanta and three Lahiri variants are all selectable per request.
03Is the interpretation generated?
Positions, dashas, divisional charts and yogas are calculated, never generated — they come from the ephemeris and a deterministic rule engine. Written interpretation is composed against those fixed facts and against a corpus of classical texts, then verified: every position it states is checked back to the chart, and every citation is checked to resolve to a real verse. If verification fails, it does not ship.
04How accurate are the calculations?
Positions derive from the Swiss Ephemeris. The engine is regression-tested against a golden dataset of documented charts cross-referenced with established Jyotish software, and a calculation regression blocks the deploy.
05Can I use the API for my own app?
Yes. Kalpa is the public API and is priced for it, from a starter tier through to enterprise and white-label. You get the same engine the other two products run on.
06Who owns the birth data I send?
You do. Birth data is treated as sensitive personal information: retained minimally, never sold, and never used to train anything without explicit consent.
Put the engine to work.
Query it directly, or use one of the products built on it.