audio_detach MCP tool: idempotency_key, jobs_wait, jobs_result

Hosted MCP exposes audio_detach. It is a write tool: it needs an idempotency_key, plus mcp:write under OAuth. The flow is audio_detach, then jobs_wait, then jobs_result, and the result carries the new audio_url.

What are the tool's inputs?

The docs describe the REST job body for POST /v1/audio-detach; the MCP tool's exact input schema is in tools_list, so read this table as the REST body from the Audio detach docs:

What does the result contain?

When the job is ready the result is kind: audio_detach with audio_url (a new artf_ artifact), duration_seconds, format, channels, sample_rate, source_duration_seconds, and optional range and warnings[]. The public rate is $0.01 per job.

Where do the files come from?

Hosted MCP cannot read files from your laptop. Upload through assets_create, assets_upload_url and assets_complete, or import with media-imports_create, before you call audio_detach.

Is there a GET for the detach?

No. There is no GET /v1/audio-detach/:id; poll the job envelope with jobs_wait or GET /v1/jobs/:id/status.

How does the job run?

Audio detach defaults to mode: "async". Pass mode: "sync" to wait up to 30 seconds for a 200 finished job, or you get 202 and poll GET /v1/jobs/:id/status and GET /v1/jobs/:id/result. There is no GET /v1/audio-detach/:id. Idempotency-Key is required. The result is kind: audio_detach with audio_url, duration_seconds, format, channels and sample_rate; sample_rate is null when you omitted it and the source rate was inherited. Under MCP, jobs_wait replaces the manual poll.

Detach is not clip inspection. For probing, stills or transcripts use video inspect; for a new MP4 cut use video trim, described in the Audio detach docs.