Luma

Luma (/media/luma) is Luma AI's Dream Machine, which turns text and images into realistic, fluid video. Its Ray models are known for natural motion, strong physics, and fast creative iteration.

Where to configure

Configure your Luma API key at Settings → Media. Postmill stores the key encrypted at rest; there is no environment-variable fallback.

See Settings for provider setup and the media capability matrix.

Tabs / operations

Tab	Operation	Key fields	Produces
Text → Video	video	Model, prompt, aspect ratio, resolution, duration, loop	A Luma Dream Machine video clip
Image → Video	video	Start frame, optional end frame, model, prompt, aspect ratio, resolution, duration, loop	A video clip animated between keyframes

Available models are Ray 2, Ray Flash 2 (faster), and Ray 1.6. Aspect ratios include 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9. Resolutions are 540p, 720p, and 1080p; durations are 5s or 9s.

Generation flow

Luma video generation is async. After submission, the job appears in the Render Queue and polls until Luma reports completed; the finished MP4 is saved to /files. From there you can open it in the Designer or post directly. See Media Studios for the shared flow.

Caveats

• Video-only: this provider does not generate still images or audio.
• Webhook with poll fallback: when your backend has a public HTTPS URL, Luma calls back on completion; otherwise (and as a fallback) completions rely on the shared media-jobs poll cron, so the queue updates on its polling interval.
• Keyframe handling: start_image_url and end_image_url are folded into Luma's nested keyframes structure by the adapter.
• Loop option: the loop toggle produces a seamlessly looping clip when enabled.

Related docs

• Media Studios — the shared render-queue/hand-off flow used by other studios.
• Designer — the target of every "Edit in Designer" hand-off.
• Settings — configuring the Luma media provider.

---

Verified against v1.3.0