Permissive license · Verifiable · Embeddable
Not a walled garden. A core.
Most astrology sites hide a lookup table or a paid API behind a pretty chart. AstralMath does the opposite: the calculation engine is open-source, so you can read it, run it, verify it, or embed it in your own product.
Why open source
In astrology, “trust us, the chart is accurate” is worth nothing — every site says it, and most are wrong by degrees because they interpolate a table or proxy a cheap API. The only credible answer to how do you know it’s right? is: read the code and run it yourself.
That is why the engine is open. Verifiability is the trust model. An engineer can audit the ephemeris calls, reproduce a position against an observatory source, and confirm there is no fudge factor. Openness is also what lets AstralMath be a shared core — a computation layer others build on — rather than one more closed app.
The stack
Every layer runs client-side. There is no calculation server to trust because there is no calculation server.
The astronomical core
Swiss Ephemeris
The high-precision ephemeris used across research and professional astrology software, derived from NASA JPL’s DE431. It resolves planetary positions to the arc-second. We run it unchanged — the astronomy is not our approximation.
Where it runs
WebAssembly
The C engine is compiled to WASM so it executes at near-native speed inside your browser. Every calculation happens on your device; your birth data never leaves it and there is no server round-trip.
The data layer
duckdb-wasm
Reference tables — nakshatra boundaries, solar terms, lots — are queried locally with DuckDB compiled to WebAssembly. Structured, fast, and still entirely client-side.
Standing on real work
AstralMath is new, but the mathematics it runs is not. We name the lineage with respect — this engine exists because of the projects below.
Swiss Ephemeris
The foundation. Astrodienst’s ephemeris is the gold standard for arc-second planetary positions, and it is what our engine compiles and runs.
Skyfield
The Python astronomy library that set the bar for readable, verifiable positional computation. A reference point for doing this transparently.
kerykeion
An open-source Python astrology library that showed how much of the chart layer can be made clean, typed, and reusable.
Roadmap: MCP for AI-native astrology
This is where the world is going. People no longer open an astrology site — they ask an AI. But large language models approximate the ephemeris and drift by degrees, confidently returning a chart that is quietly wrong.
The fix is to let the assistant call a real engine. A planned Model Context Protocol (MCP) server will expose AstralMath’s calculations as tools an AI can invoke directly — exact positions, dashas, aspects, and lots, computed by the same open core. That turns AstralMath into the computation layer behind AI-native astrology: honest math the machine can cite instead of hallucinate.
Read the math it computes
See exactly what the engine measures, what it interprets, and why we keep those two strictly apart.