A complete reference for how the site works: every prayer-time, moon, crescent, and Qibla calculation, the exact methods and formulas behind them, and every external service and data source used. Everything is computed for the location you choose.
Every figure depends on your latitude and longitude, because the Sun and Moon sit at different angles depending on where you stand. Location is resolved in three ways, in this order of accuracy:
ipapi.co service (shown as “approximate, by IP”).Prayer times come from the nojumi.org API (the Najaf Astronomical Research Center), which uses the Jafari (Najaf) method. The request is made server-side — your browser never sees the API key.
Far from the equator near midsummer, the sky never darkens enough for true dawn, so Fajr does not occur and the source returns no value. The site estimates it with the nearest-latitude rule (Aqrab al-Bilad) — using a reference latitude of 48° on your longitude. Estimated values are labelled “estimated · high latitude”.
Computed entirely on the server with the astronomy-engine library (no external call): phase & illumination, lunar age, previous/next new moon, and next full moon.
The Hijri (Islamic) date uses the Umm al-Qura calendar via the internationalization engine (ICU). The actual start of each month is a matter of crescent sighting, which the Moon Sightings tab predicts separately.
The astronomical new moon (conjunction) is when the Moon is “born” — but at that instant it is invisible. The Islamic month begins when the thin crescent is first seen, from ~15 hours to a couple of days later.
cos(ARCL) = cos(ARCV) × cos(DAZ)W = SD × (1 − cos ARCL)Every criterion begins from the same measured geometry at the observer’s location (section 6): ARCV (the Moon’s altitude minus the Sun’s), DAZ, ARCL, and the crescent width W in arcminutes. What differs is how each turns that geometry into a verdict. The site computes all three for every evening — and for every point of the world map.
Developed by B. D. Yallop at HM Nautical Almanac Office (Technical Note No. 69), as a single-number formularization of Bruin’s 1977 theoretical visibility curves, which model how bright the twilight sky is versus how much light the thin crescent delivers. It is evaluated at the “best time” (sunset + 4⁄9 of the moonset lag) and compares ARCV against a cubic polynomial in the crescent width:
q = (ARCV − (11.8371 − 6.3226·W + 0.7319·W² − 0.1018·W³)) / 10The value of q places the evening in one of six zones:
Mohammad Odeh of the Islamic Crescent Observation Project (ICOP) kept Yallop’s cubic shape but refitted its constant against roughly 700 verified modern observations — including sightings with binoculars and telescopes — so its boundaries track the real observational record rather than a theoretical sky model. Also evaluated at the best time:
V = ARCV − (7.1651 − 6.3226·W + 0.7319·W² − 0.1018·W³)Note that Odeh’s B is not a naked-eye zone — which is why the site’s first-sighting flag counts only Odeh A.
Khalid Shaukat’s criterion, used by moonsighting.com and calibrated on roughly 900 observation reports collected over decades. Unlike the other two it is evaluated at sunset and is binary: the crescent is judged visible when
S = M / 12.7 + W / 1.2 > 1 (with M ≥ 3.4°)where M is the Moon’s topocentric altitude at sunset and W the crescent width. The altitude floor rejects evenings where the Moon hangs too low in bright twilight regardless of width. Shaukat’s complete algorithm was never published; the site reproduces this documented core, and grades the map’s colour bands by margin from the S = 1 line (our own addition, to convey uncertainty — the criterion itself has no zones).
Shaukat is generally the most optimistic; Yallop the most conservative; Odeh sits in between and tracks the modern observational record most closely. The site shows all three side by side and flags the first naked-eye sighting only.
The imkān al-ruʾya criterion adopted in 2021 by MABIMS — the religious authorities of Brunei, Indonesia, Malaysia and Singapore. Unlike the three fitted models above it is a plain two-part threshold evaluated at sunset: the crescent is deemed visible when
altitude ≥ 3° AND elongation ≥ 6.4°where the altitude is the Moon’s height above the horizon at sunset and the elongation is its angular separation from the Sun. There is no crescent-width term and no “best time” — it is a deliberately simple, administrable rule for fixing one official calendar across a region. That simplicity makes it more permissive than the observation-fitted criteria: it admits many low, thin crescents that Yallop or Shaukat would reject. The site reports a plain visible / not visible verdict, and on the map shades a narrow amber band just below the line to show near-misses.
MABIMS is offered for comparison, and because it is the reference many regional calendars follow; it is not the Jafari method the rest of the site uses.
The three fitted models all answer the same question — will a human see this crescent? — but they are different models of it, and the visibility map is each model’s inequality painted across the globe. (MABIMS answers it differently again: a fixed administrative threshold rather than a fit to observations.) Four concrete differences move the curves:
First visibility is a perception threshold, not a crisp physical event, so every criterion is an empirical fit — honest disagreement, not error. In practice the three agree over most of the Earth and differ only in a marginal ribbon, which is why communities using different criteria can legitimately begin the month a day apart. The map’s All 3 view overlays them: the fill shows how many criteria call the crescent naked-eye visible, and each criterion’s own boundary is traced in its colour.
The drawn crescent shows how it will appear from your location at best time: the illuminated fraction sets the sliver thickness, and the tilt orients the bright limb toward the Sun.
The coloured world images are produced in four steps:
q and Odeh’s V at best time and Shaukat’s S at sunset, stored as numbers per cell.Selecting Shaukat, Yallop, Odeh or MABIMS paints that criterion’s zones with the legend beneath; the printable PDF is generated server-side from the same cached grid and mirrors whichever view is on screen. (MABIMS is shaded green where it meets the altitude + elongation threshold, amber just below it, and slate where clearly not visible.)
The fourth button overlays the three criteria on one image. Every point is tested against each criterion’s naked-eye rule — Shaukat S > 1, Yallop zones A–B (q > −0.014), Odeh zone A (V ≥ 5.65) — and two things are drawn:
The All 3 PDF reproduces the same fills, boundary lines, and legends from the same cached data.
Each predicted evening shows the cloud cover near sunset from the free Open-Meteo service. Forecasts only reach about 16 days ahead.
The Qibla is the direction to the Kaaba in Mecca (21.4225°N, 39.8262°E), computed as the great-circle bearing from your location, clockwise from true north. On a phone, the compass reads the orientation sensor and rotates the dial live.
As morning approaches, the sky brightens in stages defined by how far the Sun’s centre sits below the horizon. Astronomers name three twilight bands; Islamic law distinguishes two dawns. The Today page lists all of them for your location, ordered from full dark to sunrise.
Dawn is simply the morning side of twilight; the same three bands mirror at dusk.
Each astronomical band is the instant the Sun ascends through its depression angle (18°, 12°, 6°), found with the astronomy engine’s altitude search. True dawn uses the Jafari (Najaf) 16° — the same angle used for Fajr and the high-latitude estimate. The false dawn has no single agreed angle; it appears in the dark window between astronomical dawn (18°) and true dawn (16°), so the site shows it as that window rather than one clock time. Near midsummer at high latitudes the Sun may never reach these depressions, so true night — and these dawns — may not occur; such cases show as unavailable.
The Eclipses tab lists upcoming solar and lunar eclipses for your location, with the local time and the Sun/Moon’s altitude at each contact phase. Everything is computed on the server with the astronomy-engine library — no external service — so the figures agree with published eclipse tables to well under a minute.
An eclipse is a shadow alignment. Earth and the Moon each cast two nested shadows: a dark inner cone, the umbra, where the light source is completely blocked, and a faint outer cone, the penumbra, where it is only partly blocked. Which shadow is crossed, and how deeply, defines the kind of eclipse.
The gold Ṣalāt al-Āyāt window on each card is derived from these numbers: it runs from the start of the (umbral/partial) eclipse to full clearing, clipped to when the Sun/Moon is actually above your horizon — see the note on that tab.
The astronomical positions, Qibla bearing, and prayer times are computed precisely. The crescent predictions are for ideal conditions. Treat all of this as a guide, not a ruling: for religious decisions, follow your local authority.