Developer Unveils Claude-Powered Local COG Visual Explorer

Max Lenormand showcased a fully Claude-powered local Cloud Optimized GeoTIFF (COG) explorer. The tool aims to transform engineering-heavy raster file structures into intuitive visual interfaces, helping developers move beyond dry CLI metadata tables to explore geographic imagery more interactively.

Tool Capabilities and Test Subjects

Designed for local `.tif` COG files, this tool allows users to directly inspect the overview and tile structure of any file. Users can compare projection resolutions on a map and visually assess the storage footprint of each overview, spatial resolution differences, and how tiles are split. The author has processed various local COG files with the tool, testing objects including a Maxar/Vantor open data image, SLC SAR imagery, and elevation data to observe distinct visual differences in high-resolution imagery at larger scales.

Implementation and Practical Tips

Technically built using Fable and Fused Render, the tool outputs a locally running HTML + Python bundle entirely driven by Claude. The author shared practical tips for using Fable: advise making the goal highly specific in the first prompt (e.g., explicitly stating the need to visualize tiles, overviews, and storage across overviews) and letting the model attempt multiple implementations. He believes AI holds immense potential in technical data visualization, making complex engineering information easier to explore and explain.

2026-07-13 ~ 2026-07-13 · 7 related posts