Use with AI Assistants
A skill and an llms.txt that let ChatGPT, Claude and coding agents open Terrain Viewer on any place and visualization
Every setting of Terrain Viewer lives in its URL, so an assistant can answer "show me the slope around Mont Blanc in 3D" with a link that opens exactly that. Two things teach it how.
The Terrain Viewer skill
A skill is a small instruction file an agent loads when a request matches it. The terrain-viewer skill tells the agent how to pick the camera, the view mode, the visualization and the terrain source, how to open a DEM file or tile URL the user shares, and which pages hold the full reference. It lives in the repository at skills/terrain-viewer/SKILL.md.
Install it for most agents (Claude Code, Codex, Cursor, Copilot and others) with skills.sh:
npx skills add Iconem/terrain-viewerOr by hand:
- Claude Code: copy the
skills/terrain-viewerfolder into~/.claude/skills/(for every project) or.claude/skills/(for one). - Claude apps: zip the
terrain-viewerfolder and upload it under Settings, Capabilities, Skills.
Then ask in plain words, for example:
- "Show me the terrain around the Matterhorn with slope, in 3D."
- "Open this DEM in a browser: https://example.com/dem.tif"
- "Compare the IGN LiDAR nDSM with the default terrain over Paris."
- "Give me a link to spot old earthworks near Stonehenge."
llms.txt, for any assistant
Assistants that read the web, ChatGPT search included, look for /llms.txt at the site root. Ours explains how to build a link, with examples, and lists every documentation page. The complete parameter list is the URL API page and its machine-readable OpenAPI description.
Embedding & URL parameters
Everything the app shows is in its URL — how to compose a link or an iframe that loads your own COGs and vector data, the parameters that are instructions rather than state, and the project JSON preset format
Settings
The Settings dialog — appearance, caching, storage, beta features, API keys, map bounds and project presets