Developer docs

Build with the Vertical AI engineering APIs. Conversational gateways for both agents, plus focused HTTP services for parsing, classification, loading, clearance, sag, ampacity, fault current, and more.

The platform is organized as a small set of HTTP services on top of two shared engineering models: the Universal Pole Model for overhead distribution and the Universal Vault Model for underground. Parsers produce these documents, analyzers consume them, and exporters write them back out into the source format. If you're integrating with more than one API, read the Universal Pole Model reference first — every overhead service speaks it.

Most teams start with one of the conversational gateways — they orchestrate the underlying tools for you — and drop down to the individual services when they need a single piece of behavior in a larger pipeline.

All services authenticate the same way (a bearer API key) and return JSON. The gateways stream Server-Sent Events so chat UIs can show tool calls and intermediate results as they arrive.

Get started

Overhead — parsers and on-ramps

Everything that turns a source file or capture into a Universal Pole Model.

Overhead — analyzers

Every analyzer accepts a UPM document and returns the same document with findings, proposals, or computed fields enriched.

Overhead — workflows

Higher-level services that consume findings from the analyzers and produce documents, cost estimates, or portfolio plans.

Underground

Underground is split into a conversational gateway (live) and six standalone calculator APIs that are extracting from the underground agent over the next few releases. Preview pages describe what each will do.


Need access?

API keys are issued per organization. Email hello@epcstudio.io with your team and intended use case and we'll get you a sandbox key.