CalcTree calculations are now available as tools inside AI assistants. Install the CalcTree skill in Claude, ChatGPT, Gemini, Cursor, or any LLM, and it can discover your pages, run them with custom inputs, and build new ones — with real physical units and server-side computation.
What it does
The skill teaches any large language model three things: how to list the calculation pages in your workspace, how to execute any page with input overrides and read back results with units, and how to create new pages from scratch with formulas, datasets, and traffic lights.
Calculations run server-side on CalcTree's engine. The AI reads the results — it does not compute them itself, so the numbers are always exact.
Why it matters
LLMs are good at reasoning but unreliable at arithmetic. Engineering calculations need both. By giving the AI access to CalcTree as a tool, you get the LLM's ability to interpret problems, choose inputs, and explain results, backed by a deterministic calculation engine that carries real physical units through every step.
Ask it to run your beam calculator with a 12 m span and a 30 kN/m load and it calls CalcTree, gets back Mmax = 540 kN·m, and reasons over the result. No hallucinated numbers.
How to set it up
You need an API key and the skill file. Generate a key in your workspace settings, download the skill from GitHub, and follow the setup steps for your AI tool in the docs.
Claude users: upload the zip under Settings → Features → Skills and enable network access. ChatGPT and Gemini users: attach the skill document to your conversation. Cursor and Windsurf users: point at the AGENTS.md file in the calctree-skills repo.
