Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > comfyui-mediakit-toolkit

ComfyUI Extension: comfyui-mediakit-toolkit

Repo Name

comfyui-mediakit-toolkit

Author
fudanll-sys (Account age: 0 days)
Nodes
View all nodes(1)
Latest Updated
2026-11-11
Github Stars
0.0K

How to Install comfyui-mediakit-toolkit

Install this extension via the ComfyUI Manager by searching for comfyui-mediakit-toolkit
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter comfyui-mediakit-toolkit in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

comfyui-mediakit-toolkit Description

comfyui-mediakit-toolkit brings Volcengine AI MediaKit into ComfyUI workflows for video enhancement, subtitle removal, speech transcription, scene analysis, background isolation, and audio/video editing.

comfyui-mediakit-toolkit Introduction

comfyui-mediakit-toolkit is an unofficial community extension that brings Volcengine AI MediaKit (https://www.volcengine.com/docs/6448) video and audio tools into ComfyUI. Instead of leaving your node workflow to upload a video, check a processing task, and download its result, you can connect MediaKit nodes directly to other ComfyUI nodes.

You can use the toolkit to improve video quality, remove burned-in subtitles, transcribe speech, isolate a person from the background, edit video and audio, and inspect footage for scenes or highlights. It also includes an environment check to help you identify setup problems before processing media.

Before running a cloud task: You need your own Volcengine account and AI MediaKit API key. Cloud processing may incur charges. Your media is uploaded directly from your local ComfyUI environment to Volcengine; it does not pass through the extension maintainer’s server. Only upload media you have permission to process.

How comfyui-mediakit-toolkit Works

Think of a ComfyUI workflow as a chain of steps: a node loads your footage, a MediaKit node processes it, and another node saves or uses the result. For example:

Load Video → Enhance → Save Video

The toolkit accepts ComfyUI’s native VIDEO and AUDIO connections where appropriate. Some analysis nodes return text instead: ASR produces subtitles, while Scene and Highlight produce structured information that another node can inspect or visualize.

When a node uses a cloud capability, the extension handles the steps behind the scenes: it uploads the input, submits the task, checks its progress, and downloads the result. The author recommends starting with MediaKit Environment Check to check the CLI, authentication, runtime, and compatibility information before you build a larger workflow.

Choosing where an editing task runs

All nodes in MediaKit/Video Editing, plus Metadata, offer a run_mode setting:

  • auto (default): Uses local processing if your installed mediakit-cli supports that capability locally; otherwise, it falls back to a cloud task.
  • cloud: Sends the task to the cloud.

A locally completed task does not incur MediaKit cloud-processing charges. auto does not guarantee a free local run, however: a fallback to cloud may be billable. The documented CLI 0.2.2 editing capabilities support cloud processing only. If cost matters, check your planned operation before running it; the repository’s cost comparison page provides per-node estimates based on official list prices.

Understanding connections and outputs

  • VIDEO → VIDEO nodes, such as Trim or Enhance, fit between a video loader and a video saver.
  • VIDEO → AUDIO nodes, such as Extract Audio, pass sound to audio-compatible nodes rather than video nodes.
  • Text and JSON outputs contain subtitles, timelines, highlights, or media details; they are not rendered videos.
  • Multiple-input nodes have numbered sockets. Connect each video or audio item to its own socket, in the order you want it used.

For a first try, import a workflow from example_workflows, select a local video you have permission to process, and check the node settings before queueing it.

comfyui-mediakit-toolkit Features

Video improvement and cleanup

NodeWhat you can doSettings that shape the result
EnhanceImprove video quality with the standard or professional version. This is not the separate fast-enhancement capability.Choose tool_version (standard or professional), output resolution (original or 240p–4K), low/medium/high bitrate_level, and fps (0 keeps the source frame rate; a specified value can be up to 120 fps). The standard version also has a scene choice for general footage, user-generated content, short drama, AI-generated content, or old film.
Enhance UltraUse generative, diffusion-model-based enhancement and repair.Choose 720p or 1080p, low/medium/high bitrate, and fps (0 keeps the source rate; specified values support 15–120 fps).
Erase SubtitleDetect and remove subtitles embedded in the picture.Useful when subtitles are part of the video image rather than a separate subtitle track.
Erase Subtitle ProPerform more targeted subtitle or on-screen-text removal.Choose whether to erase subtitles alone or other rendered text too; prioritize picture quality or output-size control; optionally limit processing to a region of the frame with restrict_region.

Speech, text, and visual analysis

NodeWhat you getUseful choices
ASRTimestamped subtitle text from speech, plus a detailed, sanitized JSON result.Set content_type to automatic, speech, or singing; let language detect automatically or choose cmn-Hans-CN or eng-US. You can also request speaker information and confidence scores.
OCRText recognized from the video picture, with subtitle text and a detailed, sanitized JSON result.Choose Subtitle for subtitle recognition or Detailed for more information. Unlike ASR, OCR reads the image rather than listening to the soundtrack.
SceneA JSON timeline of automatically detected scene segments.Adjust segment_threshold to change cutting sensitivity; use enable_clip_fade to treat fades as separate clips. Minimum and maximum segment durations are in seconds; 0 uses the algorithm’s defaults.
HighlightJSON metadata identifying potential emotional peaks or important actions.Choose Miniseries for short dramas or Game for mini-game footage.
MetadataJSON information about the video container and its video and audio streams.No content-specific settings; its default run_mode is auto.

Background removal and picture editing

NodeWhat you can doUseful choices
Matte · PortraitIsolate a person and output video with a transparent background.Choose transparent WEBM (default) or MOV.
Matte · Green ScreenRemove a green-screen or other solid-color background, producing transparent video.Choose transparent WEBM (default) or MOV.
TrimKeep a time range from a video.Enter start_time and end_time in seconds. If the end is at or before the start, the node treats it as “to the end of the video.”
SpeedChange video playback speed.Set speed from 0.1 to 4.0; 1.0 is normal speed.
Camera MotionAdd a motion effect to footage.Choose zoom, pan-zoom, orbit-360, or bounce, and set its start and end times. An end_time of 0 extends the effect to the video’s end.
FilterApply a preset visual look.Choose spring, sunset, vivid, fair_skin, or food.
FlipMirror the picture horizontally, vertically, or both.Enable at least one flip direction.
RotateTurn the picture left 90°, right 90°, or 180°.Select the matching rotate_direction.
CropKeep a rectangular area of the picture.Set the top-left and bottom-right pixel coordinates. The bottom-right point must be farther right and lower, with at least 16 pixels of width and height.

Trim and Crop solve different problems: Trim changes when the video starts and ends; Crop changes which part of the picture remains.

Combining video, sound, and subtitles

NodeWhat you can doUseful choices
ConcatPlace videos one after another.Connect 1–100 videos to numbered sockets. Leave transitions empty for direct cuts, or supply one transition resource ID for each join.
StitchShow 2–3 videos together on one canvas.Choose horizontal for side-by-side placement or vertical for top-to-bottom placement. Unlike Concat, the clips share the screen rather than play in sequence.
SubtitleBurn subtitles into the video picture.Supply an SRT/VTT/ASS file path or URL, or a subtitles_json list with text and start/end times. Choose a font (sy_black, pm_zhengdao, or zhanku_kuaile), position (bottom_center, top_center, center, or lower_third), size, color, and text effect.
MuxCombine video with a supplied audio track.Decide whether to keep the original sound or replace it; optional audio/video synchronization settings are also available.
VolumeChange the volume of a video’s sound.Set volume from 0 (mute) to 4; 1 is unchanged.
FadeFade a video’s sound in and out.Set fade-in and fade-out durations in seconds; both default to 1.

Audio and animation

NodeWhat you can doUseful choices
Extract AudioTake the soundtrack from a video and output native ComfyUI AUDIO.Choose mp3, m4a (default), ogg, flac, or wav.
Concat AudioPlace 1–100 audio items one after another.Connect each item to a numbered AUDIO socket in playback order.
MixCombine 1–100 audio items into one track.Use this when sounds should play together, rather than one after another.
Audio SpeedChange audio playback speed from 0.1 to 4.0; 1.0 is unchanged.Choose mp3, m4a (default), ogg, flac, or wav for the output.
Audio TrimKeep a specified range of an audio item.Enter start and end times in seconds.
Fade AudioAdd an audio fade-in and fade-out.Set durations in seconds; each defaults to 1. Output defaults to mp3, with m4a, ogg, flac, and wav also available.
AnimatedExtract a selected video range as a GIF or WebP animation.Set start and end times in seconds—the end must be later than the start—and choose gif (default) or webp. It returns image frames, per-frame durations in milliseconds, and an animated_url for the encoded animation.

Animation timing matters: The IMAGE frame batch does not itself preserve GIF/WebP timing. If you need the original encoded animation, use animated_url; a cloud-hosted URL remains available only for as long as the service permits.

Checking your setup

MediaKit Environment Check reports on the MediaKit CLI, authentication, runtime, and the toolkit’s saved compatibility baseline. If the installed CLI differs from the saved version or capability descriptions, the diagnostic details can flag that difference. This makes it a useful first node to run when something does not behave as expected.

comfyui-mediakit-toolkit Models

Most toolkit choices are processing modes or presets, not downloadable models you select separately:

  • Enhance standard and professional: Two versions of the Enhance capability. Use their resolution, bitrate, and frame-rate controls to match your intended output. Scene-specific settings apply only to standard.
  • Enhance Ultra: A separate generative enhancement and repair capability based on a diffusion model. It offers 720p and 1080p output choices.
  • Highlight Miniseries and Game: Select the option matching the kind of footage you want analyzed—short-drama footage or mini-game footage.

The provided documentation does not specify additional model files or model-selection requirements for these nodes.

What's New with comfyui-mediakit-toolkit

The documented current version is 0.5.0. It includes the editing nodes described above and fixes for native ComfyUI media interfaces. In practice, that means workflows can pass video and audio through compatible native connections, while Animated returns image frames together with separate animation-timing information.

Version 0.5.0 also documents numbered, expanding inputs for multi-video and multi-audio nodes, and auto/cloud execution choices for editing nodes and Metadata. The available context does not provide a version-by-version changelog for earlier releases.

Troubleshooting comfyui-mediakit-toolkit

Start by running MediaKit Environment Check. Its diagnostic details can help narrow a problem down before you rerun a potentially billable task.

The MediaKit CLI cannot be found

  1. Check whether mediakit-cli version works in the environment that launches ComfyUI.
  2. If it works in your terminal but not in ComfyUI, check whether the ComfyUI application has access to the same system PATH. Desktop launches may not inherit terminal settings.
  3. Where appropriate, launch ComfyUI from a terminal in which the CLI is available, then run Environment Check again.

Authentication fails

  1. Confirm that your AI MediaKit API key is valid for your account.
  2. Recheck the CLI’s authentication setup, or confirm that the process launching ComfyUI receives MEDIAKIT_API_KEY.
  3. Run Environment Check again. Do not post your complete API key in screenshots, workflows, or public logs.

See the official MediaKit CLI documentation for its authentication guidance.

MediaKit nodes do not appear in ComfyUI

  1. Confirm that your ComfyUI version supports the native VIDEO API and comfy_api.latest required by the extension.
  2. Fully restart ComfyUI.
  3. Check the ComfyUI startup log for an extension import error, which may explain why the nodes were not loaded.

A cloud task times out

  1. Check the local ComfyUI log for the sanitized task ID.
  2. Do not assume that a timeout means the cloud task stopped or was not billed; it may still be running.
  3. Keep the task ID for further investigation rather than immediately submitting another paid run.

Frequently asked questions

Does auto always run locally?
No. It uses local processing only when your CLI supports that capability locally; otherwise, it submits a cloud task that may incur charges.

Does the extension send my media to the author?
No. For cloud tasks, media goes directly from your local ComfyUI environment to Volcengine. The extension maintainer does not proxy or store it.

Will my API key be saved in a shared workflow?
The extension does not write the API key into workflow data. Keep the key private and do not include it in files or screenshots you share.

Why can’t I connect an audio result to a video socket?
They are different ComfyUI media types. For example, Extract Audio produces AUDIO; connect it to an audio-compatible node. Mux can combine supplied audio with video.

Why did I get frames instead of an animated GIF or WebP file?
Animated provides a native IMAGE frame batch for ComfyUI workflows. Use its animated_url output when you need the encoded animation, and use frame_durations_ms when frame timing matters.

Learn More about comfyui-mediakit-toolkit

This is an unofficial community integration, not an official Volcengine or ComfyUI product. It is available under the MIT license.

comfyui-mediakit-toolkit Related Nodes

RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

comfyui-mediakit-toolkit detailed guide | ComfyUI