Caliper is a geometry file metadata API built for AI agents. Give any agent the ability to understand 3D files — STL, OBJ, PLY, PCD, LAS, glTF — with a single API call. No SDKs to install, no accounts to create, no geometry libraries to bundle.

This blog is where we will publish:

• Getting started guides for MCP and REST integration
• Agent integration tutorials for frameworks like LangChain, CrewAI, and Claude MCP
• Technical deep dives into file formats, metadata extraction, and the MCP protocol
• Use cases for manufacturing QA, asset pipelines, BIM, and point cloud triage
• Product updates as we add new formats and capabilities

Why a geometry metadata API?

AI agents are increasingly working with 3D data — triaging CAD files, validating manufacturing outputs, processing point clouds. But giving an agent geometry awareness today means bundling heavy libraries like Open3D or trimesh, handling format detection, and managing dependencies.

Caliper extracts the metadata your agent needs — vertex counts, bounding boxes, face statistics, format details — and returns it as structured JSON. Your agent stays lightweight; Caliper handles the geometry.

How it works

Caliper is MCP-native and accepts payment via x402 micropayments (USDC on Base). Here is a quick example using the REST API:

curl -X POST https://caliper.fit/api/auto-stats \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/model.stl"}'

The response includes format detection, vertex and face counts, bounding box dimensions, and more — everything an agent needs to make decisions about a 3D file without downloading or parsing it locally.

What is next

We are launching this blog alongside a series of getting started guides and integration tutorials. Follow along via RSS or check back regularly.

If you are building AI agents that work with geometry, we would love to hear from you. Visit caliper.fit to get started.