Metadata · 视频元信息

The most boring node in the pack, and I mean that as a compliment. Metadata · 视频元信息 takes a video and returns a structured JSON description of it - container, video stream, audio stream, the whole technical picture - and it has exactly zero business parameters. Give it a VIDEO, get back the spec sheet. That's the entire contract.

The inputs (all two of them, both boring)

• video - the native VIDEO object from your loader.
• poll_interval_seconds / max_poll_attempts - the pack's standard polling pair (defaults 10 and 720). Yes, even a metadata probe is a cloud task that gets polled.

Two STRING outputs: metadata_json - the standard metadata summary - and raw_json, the complete redacted result. Both are JSON you can wire into a display node, a script that makes decisions off the specs, or a logging step. The pack's redaction pass scrubs keys and signed URLs before you ever see them.

How it works

Same machinery as every node here: the clip materializes to a temp file, mediakit-cli submits a probe task to your MediaKit account, the node polls, then returns the result payload as the two JSON strings. It's read-only and cheap - no enhancement, no matte, nothing destructive - which makes it the least expensive thing in the pack to run.

Install and the cloud setup

Via ComfyUI Manager, search "MediaKit Toolkit" / mediakit-toolkit, fully restart - or:

cd ComfyUI/custom_nodes
git clone https://github.com/fudanll-sys/comfyui-mediakit-toolkit.git

No Python dependencies, no model downloads. You need a Volcengine account with AI MediaKit activated, an API key, and the CLI installed and authed in the environment that launches ComfyUI:

npx @volcengine/mediakit-cli install -y
mediakit-cli init --mode cloud-first --api-key "YOUR_MEDIAKIT_API_KEY" --credential-store config --yes

The honest take

Let's be direct: if you just want to know a local file's specs, ffprobe is free, instant, and doesn't send anything anywhere - the cloud probe only makes sense inside an all-cloud pipeline where everything else in your workflow is already going through MediaKit. Its value there is real, though: it lets an automated pipeline stay fully server-side, handing the metadata to a downstream node or script without a local tool dependency. In other words, it's a utility node for consistency, not for capability. Grab it when you need a uniform metadata step in a MediaKit workflow; skip it when a terminal command would do.