Skip to main content
1

Check production health

curl -sf https://vimaspatial.tech/api/health
curl -sf https://vimaspatial.tech/api/cii/summary
curl -sf https://vimaspatial.tech/mcp/health
2

Skim the submission story

For judges, teammates, or agents that need the concise product narrative, start with the Devpost submission page:
https://docs.vimaspatial.tech/docs/devpost-submission
3

Run the CLI from GitHub

Until the package is published to PyPI, install from the repository subdirectory:
uvx --from "git+https://github.com/philip-chen6/vima.git#subdirectory=packages/vima-agent" vima doctor
4

Analyze a sample frame

uvx --from "git+https://github.com/philip-chen6/vima.git#subdirectory=packages/vima-agent" vima analyze --sample masonry-p --json
5

Connect an MCP client

Point Claude Code, Cursor, Windsurf, or another MCP-capable client at:
https://vimaspatial.tech/mcp
For curl or browser checks, use https://vimaspatial.tech/mcp/health. A plain GET to /mcp can return 406 because the route expects an MCP streamable HTTP client.

MCP: Doctor

Call vima_doctor to verify the API, CII summary, and spatial zone endpoints.

MCP: CII Summary

Call vima_cii_summary for the hosted wrench-time metrics.

CLI: Skill Print

Run vima skill print --agent codex to hand off Vima context to another coding agent.

API: Frame Rows

Fetch /api/cii/frames when you need raw per-frame evidence.