---
title: "Verify alert on demand"
method: POST
path: "/api/v1/verification/ondemand"
tags: ["verification"]
---

# Verify alert on demand

`POST /api/v1/verification/ondemand`

Async on-demand verification using category-based prompt lookup. Returns HTTP 202 with a correlationId immediately. VLM processing runs in the background and results are published to Kafka / Elasticsearch via the same sink as the Kafka pipeline.

## Request body

- OnDemandVerificationRequest — Full Incident payload for on-demand verification. Accepts the same message structure that DirectMedia receives from Kafka. ``category`` and ``info`` (with ``media_urls`` + ``media_type``) are required; all other Incident fields are optional and passed through to the response.
  - `category` string, required
  - `info` object, required — Must contain media_urls (list) and media_type ('video'|'image')

## Response `202`

Verification request accepted for background processing

- object
  - `status` string, required
  - `correlationId` string, required
  - `message` string, required
  - `timestamp` string, date-time, required

## Other responses

- `400` — Unknown category or invalid request
- `422` — Validation Error

---

[API](https://skmtc.dev/nvidia/apis/alert-agent-api.md) · [All operations](https://skmtc.dev/nvidia/apis/alert-agent-api/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/nvidia/alert-agent-api/revisions/e754ab3efedd/schema)