@-mention an agent on a video asset: Runway vs Sume

Runway's changelog for Sep 15, 2026 says Enterprise users can @-mention the Agent directly in comments on any asset page. Sume does not have a comment-thread trigger in its docs; you hand a Sume Agent a task through the Agent Completions API or a thread, and media jobs report back by webhook.

Two ways to hand work to an agent

According to the Runway changelog, the entry point is a comment on an asset page, limited to the Enterprise plan. In Sume the entry point is a request from your own backend or a thread in the Agents chat.

The Sume path in detail

An Agent Completion runs the Sume Agent on an ad-hoc prompt with the same runtime as the Agents chat UI: sandbox, tools, MCP bridge and media generation. The create call returns 202 with a receipt, and you poll it, because a real agent turn takes longer than an HTTP request should stay open.

Scopes are agent_completions:read and agent_completions:write. Keys created before the feature shipped do not carry them, and scopes cannot be added to an existing key, so create a new key. generation_spend_cap_usd has no default and the request fails without it.

Getting media results back

Video generation is asynchronous. On POST /v1/videos you can pass callback_url, and Sume POSTs a signed job event when the job reaches a terminal state. Verify the x-sume-webhook-signature header, return a 2xx after storing the event, and keep polling GET /v1/jobs/{id}/status as a backup.

Which to use

If your team reviews video on a Runway asset page and wants to ping an agent there, that is a Runway Enterprise feature. If the work should start from a ticket, a queue or a customer action in your own product, call the Sume completions endpoint from the server and post the result into whatever comment system you use. Sume does not read comments for you.