Get 1080P Video

This is a follow-up node. It doesn't generate anything from a prompt - it takes the task ID of a video you already made and fetches the 1080p version of it. Once you know that, the whole card makes sense: one real input, one video out.

Why KIE splits it this way

Rendering a higher-resolution pass is a separate, slower job on the provider's side, so KIE exposes it as its own operation instead of making every generation pay for it. That's why you generate at a cheaper resolution, look at the result, and only then spend on the 1080p pass of the take you actually like. It also means you can fetch a higher-resolution file for a job generated in an entirely different session - as long as you kept the task ID.

Inputs

taskId (required) - the task ID of the Veo job. This is the whole node. The field's default is a placeholder string, not a real ID, so paste the real one; if you're chaining in the workflow, wire it from the generation node's task_id output.

index (optional) - which video to fetch, as an integer, default 0. Relevant when a job produced more than one result and you want a specific one rather than the first.

timeout_seconds - 1200 default, 30–7200. Worth raising: an upscale pass takes longer than the initial prompt-to-video call.

callback_url - leave empty for normal ComfyUI use.

Outputs

video - a real ComfyUI VIDEO object, so it goes straight into Save Video, a preview, or a later stage. url - the primary hosted result. all_urls_json - every URL the response reported; keep it, since these jobs can return companion stills alongside the video. task_id - the job handle for the fetch. raw_json - the full response, which is where you look if it returns metadata but no usable file. credits_consumed and credits_left - floats, plus the credits receipt the node draws on itself after a successful run.

Install

ComfyUI Manager can find the pack by title, or:

cd ComfyUI/custom_nodes
git clone https://github.com/felipederosilva/ComfyUI-KIE-Nodes-Next
pip install -r ComfyUI-KIE-Nodes-Next/requirements.txt

Restart afterwards. Dependencies are requests, Pillow, numpy, PyYAML - the node is a thin API client, there's nothing to download or run locally, and the work is metered per call.

Key: Settings → KIE.ai Nodes Next → Connection → Paste / replace KIE.ai API key. The pack checks it against KIE before replacing a working key and shows a masked fingerprint plus CONNECTED. Servers can use KIE_API_KEY instead, which takes priority.

Gotchas

Job-not-ready is the normal first answer. A higher-resolution pass needs time after the original generation finishes, so a fetch that comes back with metadata instead of a file usually means "ask again later." Raise the timeout, or queue this node as its own step rather than wiring it in immediately behind the generator.

The task ID has to be the right one. A generation task ID from a different model family won't resolve here, and the node will spend a call finding that out.

Keep your task IDs. There's no lookup-by-prompt and no history browser on the card. If you want to re-fetch a 1080p pass next week, the ID is the only handle - the task_id output from the generation node is worth saving into a text file in the same run.

Cards are generated from KIE's live catalogue. The pack indexes KIE's API documentation on startup and re-checks roughly daily, so field changes arrive with a ComfyUI restart rather than a plugin update.