Introduction

Panchang and its five limbs, kundli and the divisional charts D2 to D60, thirty dasha systems, yogas and doshas, muhurat windows, ashtakavarga, planetary strength, horoscope matching, numerology in thirteen systems, and eleven regional calendars with their festivals.

Everything is calculated for the date, time and coordinates in your request. There is no cached almanac: almost every endpoint returns numbers and classical classifications, not written readings. A few return written text instead: the chat endpoint, general predictions, rashifal and the PDF reports.

Most of the API behaves the way you would expect. These three do not, and they are the ones that cost people time.

Many endpoints take a keys array naming which calculations you want. One call to /api/astro/panchanga/ can return any of 57 different results depending on what you ask for.

All of them cost the same single credit. Asking for everything you need in one call is the cheap path, not the expensive one — the mistake is calling the same endpoint five times for five keys. Every endpoint page lists its valid keys, and the reference is searchable by calculation as well as by route, so you can find the feature without knowing which endpoint carries it.

{
  "date_time": "2026-09-16T06:00:00+05:30",
  "latitude": 26.9124,
  "longitude": 75.7873,
  "timezone_as_float": 5.5,
  "keys": ["tithi", "nakshatra_segments", "yogam", "karana", "vaara"]
}

Every JSON response has the same four fields. Branch on is_error, read data, and show message when something went wrong — it names the fields at fault rather than saying “invalid request”.

{
  "data":     { ... },        // the result; shape varies by endpoint
  "status":   200,
  "is_error": false,
  "message":  "successful"
}

One credit per successful call; the two PDF reports and the chat endpoint cost 25. A 4xx, a 5xx or a timeout costs nothing — the credit is reserved before the calculation runs and released if it does not return cleanly. A bad integration cannot drain your balance while you are debugging it.

Access is by a monthly or yearly plan. Its quota resets every month, and calls you did not use do not carry over. There is no per-seat charge, and you can cancel any time. Every response carries X-Credits-Charged and X-Credits-Remaining so you can watch the balance without a second call.

Positions come from NASA JPL ephemeris data — the full data set, 735 fixed stars, tested from 500 to 2999 AD. Median server-side response time is 62 ms, measured across every published example rather than a chosen one; the slow endpoints are deep scans and are documented as such.

Indian traditions genuinely disagree about the ayanamsa, so the API exposes 48 of them and lets you correct any by degree, minute and second. That choice is yours to make and to explain to your users — we do not make it for you by shipping one and calling it correct.

It is calculation, not counsel. Nothing it returns is medical, legal or financial advice, and apart from the few endpoints that return written text it writes no interpretive prose. If your product makes suggestions on the back of it, that is your editorial decision — see the licence for what you may do with the output.

The base URL is https://api.occultapi.com. Beyond plain HTTP there is a playground that runs real calls with your own key (a call that fails costs nothing), a Postman collection, an OpenAPI 3.1 spec for generating a client, and an MCP server so an AI assistant can call the API directly.