API reference

Overview

The Lirovo REST API turns a source video into typed, evidence-linked JSON. Every resource (schemas, destinations, jobs, and their artifacts) is scoped to your tenant. The base URL is https://api.lirovo.ai and every path is prefixed with /v1.

Base URL#

All requests go to https://api.lirovo.ai, and every path is prefixed with /v1. Per-PR preview deploys are served at https://preview-{pr}.api.lirovo.ai. Authenticate every request with an API key as a bearer token; the only unauthenticated endpoint is GET /v1/health.

bash

curl https://api.lirovo.ai/v1/health

The surface#

Lirovo is the structured-extraction layer for video. You register a schema (a JSON Schema describing the typed output you want), submit a job against a source video, and read back a result: the typed JSON plus the evidence that anchors every extracted value to a source moment with a timestamp and a modality (audio or visual). Pass A also builds a compact knowledge graph of nodes and edges, and the run leaves queryable artifacts behind: the transcript, the kept-frame inventory, and per-frame vision descriptions.

The same surface is also exposed as an MCP server at https://api.lirovo.ai/mcp over Streamable HTTP. These pages document the REST endpoints; the schema source of truth is schemas/openapi.yaml.

Tenant isolation

Every read and write is scoped to the tenant that owns your API key. The API never leaks cross-tenant existence: an unknown id and another tenant's id return the same 404 envelope.