Get started

Quickstart

The REST flow is three calls: register a schema, submit a job against a video, then fetch the typed result.

Three REST calls#

All requests go to https://api.lirovo.ai and carry Authorization: Bearer $LIROVO_API_KEY. Keys are stored only as a one-way hash, so copy yours when it is provisioned.

Register a schema

A schema is a JSON Schema describing the typed output you want. Post it once and reuse its id across jobs. Returns 201 with the schema id.

bash

curl -X POST https://api.lirovo.ai/v1/schemas \
  -H "Authorization: Bearer $LIROVO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "name": "meeting-recap",
        "json_schema": { "type": "object",
          "properties": {
            "summary":   { "type": "string" },
            "decisions": { "type": "array" } } } }'
# -> 201 { "id": "sch_a1b2c3" }

Submit a job

Point a job at a source_uri and the schema_id you just created. The job is accepted and queued; you get back 202 with a job id and an initial status.

bash

curl -X POST https://api.lirovo.ai/v1/jobs \
  -H "Authorization: Bearer $LIROVO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "source_uri": "https://youtube.com/watch?v=...",
        "schema_id":  "sch_a1b2c3" }'
# -> 202 { "id": "job_x9k2m4", "status": "queued" }

Fetch the typed result

Poll the result endpoint. Once the job is done you get 200 with JSON that conforms to your schema.

bash

curl https://api.lirovo.ai/v1/jobs/job_x9k2m4/result \
  -H "Authorization: Bearer $LIROVO_API_KEY"
# -> 200 { "summary": "...", "decisions": ["..."] }

On every response

Each response carries an x-lirovo-request-id header for support and tracing. See the API reference for the full request and response shapes.

Next steps#

Prefer typed calls or want an agent to drive the loop? Use the TypeScript SDK for the typed client, or connect over MCP so any MCP client can call extraction as a set of tools.