Docs

Concepts to read first

These hold whatever the app ends up looking like.

• Frame receipt formatHow one telescope frame becomes a receipt: a record of what was measured on it, and how receipts roll up into sessions and a star database.
• How calibration worksHow bias, dark and flat frames take the camera back off a raw frame.
• How claims are markedWhat measured, implemented, planned and the other markers on this site mean.

Reference for the 1.3 old engine

These documents live in the public GitHub repository of SkyCruncher 1.3, the old engine. They describe how that engine works, not the app in development, which is built on a new engine. Each opens on GitHub.

Running it without the desktop app

• Headless APIRun the same pipeline from a Node program and get the same receipt back.
• MCP serverLets an AI agent run a solve and read the receipt, locally on your computer.

How it works

• Processing flowEvery stage a frame goes through, from raw file to receipt.
• Location and data privacy designA design note for a hosted results archive. It isn't a description of what this site does today; that's on Privacy.
• Third-party components and licensesWhat the 1.3 app and its installer ship, and the license of each.

Subject cards

One card per part of the old engine, with the formulas checked against its code. 13 cards, as of 20 October 2026.

• Astrometry and WCSPlate solving (working out exactly where in the sky a frame points from the stars in it) and the mapping from pixel to sky.
• Quad codes and rotation sweepHow the plate solver matches patterns of four stars against the catalog.
• Photometry and statisticsHow a blob of pixels is judged a star, measured, and used to confirm a solve.
• Color and photometric calibrationFrom raw sensor counts to calibrated color and brightness, tied to Gaia star colors.
• Image and signal conditioningTurning raw sensor data into color, estimating the sky background, and the display stretch.
• Atmosphere and opticsWhat the air and the telescope do to starlight before it reaches the sensor.
• Light pollution and scattered moonlightModeling city glow and deciding when moonlight spoils a frame.
• Background maskingWhich pixels the sky-background model may trust, without machine learning.
• Terrain mappingFinding the horizon in a frame, and refusing when there is none to find.
• Time and ephemerisClocks, sidereal time and planet positions.
• Star catalog storageHow Gaia star data is split into sky cells, fetched and checked.
• Stacking and drizzleA placeholder: the card says it hasn't been written yet.
• Future equationsEquations kept for later. The card says none of them runs anywhere.

Coming with the app

planned These sections will be written when the app ships.

• Install and requirements
• First run
• Star data
• Supported files
• Outputs
• The receipt schema
• Privacy and sharing
• Troubleshooting