Nodes/comfyui_lumaAPI/LUMA_YoC_result @.@
ComfyUI Node

LUMA_YoC_result @.@

The node that hands you the finished Luma video (eventually)

By superyoman·Created 2 years ago·Updated 2 years ago· 21
LUMA_YoC_result @.@
    • video_link
    textEnter your id
    downloadtrue

    This is the second half of the comfyui_lumaAPI pack and the one people forget to plan for. The first node, LUMA_API_YoC, fires off a Luma Dream Machine generation and hands you a string id. This one takes that id, waits for Luma's cloud to finish rendering, and returns the video link - downloading the file to disk for you if you want. Two nodes, two workflows, one cloud account.

    What it's actually for

    Luma generations aren't instant. You submit, the queue grinds, and minutes later there's a video. LUMA_API_result_YoC exists to close that loop: it polls your account for the generation, matches your id, and hands you the finished video_link. The pack's example workflow even uses a third-party ShowText (pysssss) node to display the link on screen, which is worth stealing - there's no built-in viewer.

    How it works

    Under the hood it's refreshingly simple. The node calls Luma's internal /user/generations/ endpoint (offset 0, limit 10), scans the returned list for your id, and when it finds a match pulls the video URL. If download is on, it streams the file to disk as output/luma video/<your-id>.mp4. You get a STRING back either way - that's your video_link output. Wire it into a ShowText node, an "open URL" node, or your own file handling.

    The inputs that matter

    • text - the generation id from LUMA_API_YoC, or from record_id.txt in the output folder. The default "Enter your id" is not a hint; it's a placeholder that will return nothing useful.
    • download - boolean, default true. Off if you just want the link; on if you want the mp4 saved locally.

    Output is a single video_link STRING.

    Installing it

    Same pack, same steps as LUMA_API_YoC - ComfyUI Manager, or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/superyoman/comfyui_lumaAPI.git
    # restart ComfyUI
    

    No models to download. The only real prerequisite is a valid cookie in the pack's key.json from a logged-in lumalabs.ai session - the file ships with a dead one, so replace it before anything works.

    Where people get burned

    Three things bite. First, "No result" is a real string, not an error. If the id isn't in your last 10 generations yet - or ever - the node happily outputs the literal text "No result" and your graph keeps running like nothing's wrong. If you only look at the final video file, you'll wonder where it went.

    Second, the 10-generation window. The endpoint returns only your ten most recent generations. Submit more than nine videos, then come back for an older one, and the node can't see it - "No result" again, even though the video exists. Grab the id and check promptly.

    Third, the save path is relative to wherever you launched ComfyUI from. Start it from inside the ComfyUI folder and the download lands in ComfyUI/ComfyUI/output/luma video, a nested directory you'll swear doesn't exist until you find it. Also expect Luma's watermark on free-tier output - the pack's own example link ends in watermarked_video...mp4.

    The usual pack caveats apply: it rides on your session cookie, Luma can break it anytime, and the README flags it as learning-only. But as a free bridge from a ComfyUI graph to genuine cloud video, it does exactly what it says on the tin.

    CategoryYoC app/LumaAPP

    Inputs (2)

    NameTypeDefaultDescription
    textSTRINGEnter your id
    downloadBOOLEANtrue

    Outputs (1)

    NameTypeDescription
    video_linkSTRING