Status: Planner, capture, and rendering are functional. Tested with real hardware (LX200 OnStep + Canon 600D) and rendered video output with star-aligned linear pan and crossfade transitions.

A browser-based application for planning and executing imaging sequences with a remote-controlled telescope. Draw a path on an interactive star map, and the telescope follows it point by point, capturing images at each position. A companion rendering app converts the captured FITS frames into video.

• Plan imaging sequences by drawing spline paths on a live star map (Stellarium Web Engine)
• Capture images by controlling your telescope and camera via INDI protocol over the network
• Export sequences for EKOS/KStars as an alternative to direct INDI control
• Render captured frames into video with debayering, stretch, star alignment, and transitions

git clone [email protected]:zworkb/AstroNightcrawler.git
cd AstroNightcrawler
make install    # Python venv + all dependencies
make run        # Start the planner app (auto-downloads sky data on first run)

Open http://localhost:8090 in your browser.

# CLI mode — render a capture directory to video
nightcrawler-render --input ./output --output video.mp4 --transition linear-pan

# Web UI mode — browser-based preview and rendering
nightcrawler-render --ui

The renderer reads the manifest.json written during capture. Each captured frame is debayered, stretched to 8-bit sRGB, optionally aligned via star triangles, and assembled into video with transitions.

The star map requires a pre-built Stellarium WASM binary. Either:

• Copy static/stellarium/ from an existing installation, or
• Build from source: ./scripts/build_stellarium.sh (requires Emscripten, ~10 min)

Without it, the planner works but shows no star map.

• Python 3.11+
• uv (recommended) or pip
• ffmpeg (for video rendering)
• Optional: INDI server + telescope/camera for actual capture

All settings via environment variables or .env file (created automatically from .env.example on first make run):

Two separate applications connected by a self-describing data format:

Planner & Capture App (nightcrawler) — runs on the telescope control machine or any machine with network access to the INDI server

• Interactive star map with offline star catalogues (Stellarium Web Engine)
• Cubic Bezier spline path editor with configurable capture point spacing
• Pure-Python async INDI client with BLOB support (replaces PyIndi for reliable network operation)
• Capture controller with pause/resume/cancel, automatic retry, and safety abort
• NiceGUI web interface served via FastAPI/uvicorn

Rendering App (nightcrawler-render) — runs on a powerful workstation

• FITS import via manifest.json (shared Project model)
• Bayer demosaicing via colour-demosaicing (auto-detects pattern from FITS header)
• Tone mapping: ZScale+Asinh (auto), histogram percentile, or manual black/white/midtone
• Star-based frame alignment via astroalign (with optional downsampling for speed)
• Transitions: crossfade blending or linear-pan with sub-pixel shifting (scipy)
• Resolution presets: native, 4K, 1440p, 1080p, 720p
• H.264 video encoding via ffmpeg
• Both CLI and web UI interfaces

The two apps communicate through a directory of FITS files plus a JSON manifest (manifest.json), which is the Project model serialized after capture.

make install         # Install everything
make run             # Start the planner app
make test            # Run tests
make check           # Linting (ruff)
make mypy            # Type checking

• Design Specification
• Async INDI Client
• Architecture (UML)
• Dependencies
• Stellarium Build Guide
• Project Board

• Telescope: LX200 OnStep (German Equatorial Mount)
• Camera: Canon DSLR EOS 600D via gphoto2/INDI
• Server: INDI on Raspberry Pi (StellarMate), remote network access
• Images: 5184x3456 px, 16-bit FITS (35 MB per frame)
• Rendered output: H.264 video with linear-pan transitions (star-aligned sub-pixel panning)

TBD