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 > FunUI > FunUI Upscale Video

ComfyUI Node: FunUI Upscale Video

Class Name

FunUIUpscaleVideo

Category
FunUI/Output
Author
digital-garbage (Account age: 1276days)
Extension
FunUI
Latest Updated
2026-11-03
Github Stars
0.02K

How to Install FunUI

Install this extension via the ComfyUI Manager by searching for FunUI
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter FunUI 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

FunUI Upscale Video Description

Upscales finished videos using local image models or NVIDIA RTX Super Resolution, preserving audio.

FunUI Upscale Video:

FunUIUpscaleVideo enlarges a finished video using a locally installed image-upscaling model or, when enabled, NVIDIA RTX Super Resolution. It is designed for improving a completed render rather than changing the video-generation process, so you can upscale a selected clip without generating it again. The model-based method processes the video as individual images using ComfyUI’s existing upscaling tools, including tiled processing and memory-management support. Frames are processed in small batches rather than loading the entire clip into memory, making longer videos more manageable, although larger frames and stronger enlargement still require substantial processing resources. The resulting video uses the source’s reported frame rate and includes audio when present; this is spatial upscaling, not frame interpolation. Visual results depend on the selected upscaling method and the quality of the original footage.

FunUI Upscale Video Input Parameters:

filename

Required STRING identifying the existing video file you want to upscale. The node resolves this name together with subfolder and type to locate the video in ComfyUI’s output or temporary storage. Supply the filename of an actual completed video, including its extension, rather than a descriptive name for the new result. The default is an empty string, "", which must be replaced with a valid filename for processing to succeed. No minimum or maximum string length is specified. If the resolved file does not exist, processing stops before loading an upscaler.

subfolder

Required STRING specifying the video’s subfolder within the storage location selected by type. For example, if your video is stored inside a renders folder under ComfyUI’s output directory, use renders here and put the video’s filename in filename. The default is an empty string, "", appropriate when the video is directly inside the selected storage directory rather than a subfolder. No minimum or maximum string length is specified. This parameter locates the source video; it does not choose where the upscaled result will be saved.

type

Required COMBO selecting which ComfyUI storage location contains the source video. The exact options are output and temp, and the default is output. Choose output for a video saved in ComfyUI’s output directory, or temp for a video currently stored in its temporary directory. This choice affects source-file lookup, not the destination of the finished upscale. Selecting the wrong location causes a missing-file error even if a video with that filename exists elsewhere.

upscale_model

Required COMBO selecting the upscaling method. Model choices come from ComfyUI’s models/upscale_models location, and the interface recommends using .safetensors files. When the nvidia_vsr add-on is enabled, the list also includes the NVIDIA RTX Super Resolution option. No explicit default is declared in the schema; if no model choices are found and the RTX option is unavailable, the list falls back to a single empty entry, "", which is not a usable model. For the model-based route, the selected model determines the enlargement behavior and influences sharpness, texture, processing time, and memory use; the node does not expose a separate model-scale control. For the RTX route, configuration is supplied through rtx_options. There is no fixed numerical minimum or maximum because this parameter selects an available method rather than a numeric value.

rtx_options

Optional STRING containing a JSON object for NVIDIA RTX Super Resolution settings. The default is "`{}`", an empty JSON object that leaves configuration to the RTX options handler. The documented setting names are quality, resize, scale, width, height, strength, and tenbit. These cover quality, resizing, dimensions, enhancement strength, and a ten-bit option, but their accepted values, units, ranges, and individual defaults are not shown in the supplied code and should not be assumed. Use the RTX add-on’s documented settings when changing them, and keep the text valid JSON. This parameter is only parsed when NVIDIA RTX Super Resolution is selected; it does not configure ordinary upscale models. No minimum or maximum string length is specified.

FunUI Upscale Video Output Parameters:

This node has no output sockets and therefore returns no named output parameters for connecting to other nodes. Instead, it saves the processed video in the funui_upscaled subfolder of ComfyUI’s output directory and returns video information to the interface for display. The video is encoded as an H.264 MP4 using quality setting CRF 16 and the widely supported yuv420p pixel format. Its frame rate is taken from the source’s reported frame rate, and, when audio exists, the first audio stream is included and re-encoded as AAC at 192 kb/s rather than copied unchanged. The encoding step rounds each output dimension down to an even number when necessary, which can reduce an odd width or height by one pixel.

FunUI Upscale Video Usage Tips:

  • Check the complete source location before running: filename should identify the video, subfolder should match its location beneath the selected storage directory, and type should be either output or temp as appropriate.
  • Install compatible upscale models in ComfyUI/models/upscale_models and confirm that your intended model appears in upscale_model. The interface recommends .safetensors files; an empty selection is not a working upscaler.
  • Test a short representative clip before processing a long render. Include motion, faces, fine textures, or other important details so you can judge whether the selected model produces suitable results across frames.
  • Choose a model with an enlargement factor appropriate to your delivery size. A larger enlargement can substantially increase processing time, memory use, and file size, and this node does not expose a separate scale adjustment for the model-based route.
  • Start with rtx_options set to "`{}`" when testing NVIDIA RTX Super Resolution. Only add settings supported by your installed RTX integration; the supplied node code does not establish valid ranges or example values for those settings.
  • Keep both ffmpeg and ffprobe available on the system PATH used by ComfyUI. The node needs them to inspect the source, read its frames, and encode the processed video.
  • Allow sufficient graphics memory and disk space. Frame streaming reduces the need to hold the entire video in memory, but a single high-resolution upscaled frame can still be expensive, and the completed MP4 needs space in ComfyUI’s output directory.
  • Use this node for resolution enhancement, not smoother motion or a different frame rate. The encoding follows the source’s reported frame rate; there is no frame-rate or interpolation control.
  • Expect audio to remain present rather than bit-for-bit identical. The first source audio stream is re-encoded to AAC, so additional audio tracks and the original audio encoding are not preserved by the shown encoding command.

FunUI Upscale Video Common Errors and Solutions:

[filename] is not in the [type] folder (anymore).

  • Explanation: The node could not resolve the source video to an existing file. The filename, subfolder, or storage type may be incorrect, or the source may have been moved or deleted.
  • Solution: Confirm that the video still exists, enter its exact filename including the extension, check subfolder, and select the matching output or temp location. If a temporary clip has disappeared, recreate or save it before upscaling.

ffprobe was not found on PATH: install ffmpeg to upscale videos.

  • Explanation: The video-inspection tool is unavailable to the ComfyUI process, so the node cannot read the source dimensions, frame rate, or audio information.
  • Solution: Install an FFmpeg distribution containing ffprobe, add its executable directory to your system PATH, and restart ComfyUI so it receives the updated environment.

no video stream in [filename]

  • Explanation: The inspected file does not contain a recognizable video stream. An audio-only file, an incorrect file selection, or a damaged media file can cause this error.
  • Solution: Open the file in a media player to confirm it contains playable video. Select the correct file or export a fresh video from your rendering or editing application.

no frames could be read from [filename]

  • Explanation: The processing loop received no complete video frames, so there was nothing to upscale or encode. The source may be empty, damaged, or unreadable by the installed FFmpeg.
  • Solution: Check that the source plays and has a nonzero duration. Try exporting it again in a commonly supported video format, and check the ComfyUI console for decoding errors.

ffmpeg could not encode the upscaled video: [FFmpeg error details]

  • Explanation: FFmpeg failed while creating the finished video. The message includes the end of FFmpeg’s error report, which may identify an unavailable encoder, an audio problem, insufficient disk space, or another encoding issue.
  • Solution: Read the appended error details first. Confirm that your FFmpeg installation supports libx264 and AAC, check free space and write permissions in the output directory, and test with a short, known-good source clip.

FileNotFoundError referring to ffmpeg

  • Explanation: This is a possible system-level error when Python cannot launch the FFmpeg executable; its exact wording depends on your operating system. Installing FFprobe alone does not guarantee that FFmpeg itself is available.
  • Solution: Install the complete FFmpeg tools, verify that both ffmpeg and ffprobe are accessible from the environment running ComfyUI, and restart ComfyUI after changing PATH.

Upscale model loading error

  • Explanation: The model loader may fail if upscale_model is empty, the selected model has been removed, or the file is incompatible or damaged. The supplied code does not define a single fixed error message for these cases.
  • Solution: Select an existing, compatible image-upscaling model from ComfyUI/models/upscale_models. If necessary, replace the model with a verified download and refresh the available model list or restart ComfyUI.

RTX options parsing or configuration error

  • Explanation: When the RTX method is selected, rtx_options is parsed and validated by the RTX integration. Invalid JSON or unsupported setting values can prevent processing, but the exact validation messages are not included in the supplied code.
  • Solution: Reset rtx_options to "`{}`" and test again. When adding settings, use valid JSON with double-quoted property names and follow the installed RTX add-on’s documentation for accepted values.

Out-of-memory error during upscaling

  • Explanation: Model-based processing uses ComfyUI’s tiling and out-of-memory back-off mechanisms, but large frames or demanding models can still exceed available memory. Small frame batches do not eliminate the memory required to upscale each frame.
  • Solution: Close other applications using graphics memory, try a less demanding model or a smaller enlargement factor, or prepare a lower-resolution source. There is no exposed batch-size or tile-size input on this node.

Permission error while creating or writing the output

  • Explanation: The node creates and writes files inside funui_upscaled under ComfyUI’s output directory. Processing can fail if the account running ComfyUI lacks write access or the destination storage is unavailable; the exact message depends on the operating system.
  • Solution: Check that the output directory is writable, ensure the storage location is connected and has free space, and correct permissions for the account running ComfyUI.

FunUI Upscale Video Related Nodes

Go back to the extension to check out more related nodes.
FunUI
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.