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 > H3 Continuum > H3 Continuum Finalize

ComfyUI Node: H3 Continuum Finalize

Class Name

H3ContinuumAssembleSeamV35

Category
MiniMax H3/Continuum
Author
ukr8b3g-cmyk (Account age: 400days)
Extension
H3 Continuum
Latest Updated
2026-11-04
Github Stars
0.31K

How to Install H3 Continuum

Install this extension via the ComfyUI Manager by searching for H3 Continuum
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter H3 Continuum 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

H3 Continuum Finalize Description

Assembles Continuum V3.5 video and audio chunks with configurable seam correction and storage.

H3 Continuum Finalize:

H3ContinuumAssembleSeamV35 combines decoded video chunks and their corresponding audio into one final animation while applying the selected audio and video seam behavior. It is designed for Continuum V3.5 workflows and uses direct-write image assembly, which avoids an additional final image copy when Exact Duration processing is enabled. You can choose how the final image data is stored: automatically in memory when sufficient physical memory is available, explicitly in RAM, or in a disk-backed buffer when memory usage needs to be reduced. The node can preserve and use original Driving Audio from the assembly plan or an optional direct Driving Audio connection, allowing generated audio and Audio Seam processing to be bypassed when a source recording should remain unchanged. Its Video Seam options support automatic correction, a second automatic mode, analysis without correction, or no correction, while the report provides information about the completed assembly and, when detailed diagnostics are enabled, projected buffer usage and processing details.

H3 Continuum Finalize Input Parameters:

images

Connects the decoded image or video chunks that should be assembled into the final video. The node processes these chunks according to the connected assembly_plan, including their intended order and duration. The image content and chunk structure directly determine the visual result, frame count, and storage requirements. This parameter has no user-selectable minimum, maximum, or default value in the provided interface; it must receive an IMAGE value.

audio

Connects the generated or decoded audio chunks associated with the image chunks. The node assembles this audio and applies the selected Audio Seam behavior unless a Driving Audio source is selected, in which case the preserved Driving Audio takes precedence and generated audio and Audio Seam processing are bypassed. This parameter has no user-selectable minimum, maximum, or default value in the provided interface; it must receive an AUDIO value.

assembly_plan

Provides the Continuum assembly instructions used to interpret and combine the image and audio chunks. The plan supplies information such as the target frame arrangement and can contain preserved Driving Audio for later selection. It is also used when Exact Duration is enabled to determine the requested final frame count and whether the final frame should be preserved. This parameter has no user-selectable minimum, maximum, or default value in the provided interface; it must receive an H3_CONTINUUM_ASSEMBLY_PLAN value containing a single assembly plan.

exact_total_duration

Controls whether the node enforces the exact duration requested by the assembly plan. When enabled, the assembled image and selected audio are adjusted to the plan's target frame count, with the plan's final-frame preservation setting respected. This is useful when the output must match a precise duration or remain synchronized with a downstream timeline. When disabled, the node does not perform this exact-duration enforcement step. The interface identifies this as a Boolean control, but no explicit default value is provided in the supplied node interface.

audio_seam

Controls whether audio transitions between assembled chunks receive seam processing. The available options are Off and Auto. Choose Auto when chunk boundaries may contain audible clicks, gaps, or level changes and you want the node to handle those transitions automatically. Choose Off when the audio is already prepared, when exact original audio must be preserved, or when you want to avoid audio seam processing. When a valid Driving Audio source is selected, generated audio and Audio Seam are bypassed regardless of this setting. No separate minimum, maximum, or default value applies because this is an option selector.

video_seam

Controls how visual transitions between decoded chunks are handled. Auto applies the standard automatic video seam behavior, Auto 2 selects the alternate automatic mode available in this node, Analyze Only evaluates seam conditions without applying visual correction, and Off disables video seam processing. Use an automatic mode when visible discontinuities occur at chunk boundaries, Analyze Only when you want diagnostic information before changing the frames, and Off when the chunks are already visually continuous. No separate minimum, maximum, or default value applies because this is an option selector.

buffer_backend

Controls where the final IMAGE data is stored while the node assembles the result. Auto keeps outputs up to 4 GiB in RAM only when physical memory has an appropriate safety reserve; otherwise, it maps the final image buffer to disk. RAM forces memory-based storage and can be fast, but it may require substantial available system memory. Disk-backed stores the final image buffer using disk-backed mapping, which can reduce RAM pressure for large outputs but depends on sufficient disk space and may be slower. The available options are Auto, RAM, and Disk-backed, and the default is Auto.

diagnostics

Selects how much diagnostic information is included in the report. Basic provides normal assembly status information, Detailed Report enables expanded diagnostics such as projected image and audio buffer usage and the V3.5 direct-write behavior, and Off suppresses optional diagnostics. Detailed reporting can help you understand memory requirements and investigate seam or duration behavior, while Off can keep reports shorter. The available options are Basic, Detailed Report, and Off; no explicit default value is provided in the supplied interface.

driving_audio

Optionally supplies an original audio source that should be used instead of generated audio. When connected and valid, the node copies the waveform to CPU memory, preserves its sample rate, slices it for review timing when required by the assembly plan, and returns it as the final audio. The report identifies this as audio selected from the direct input unless the assembly plan already contains preserved Driving Audio, in which case the plan source takes priority. When Driving Audio is selected, generated audio and Audio Seam processing are bypassed. This optional parameter has no user-selectable minimum, maximum, or default value; it accepts an AUDIO value.

H3 Continuum Finalize Output Parameters:

images

Returns the completed assembled image sequence as an IMAGE value. This is the final visual result after chunk assembly, optional exact-duration enforcement, and the selected Video Seam behavior. The image data may have been assembled through RAM or disk-backed storage according to buffer_backend, but the output is presented to downstream nodes as the final image sequence. Its frame count should correspond to the assembly plan, especially when exact_total_duration is enabled.

audio

Returns the audio associated with the assembled images as an AUDIO value. Normally this is the assembled chunk audio after the selected Audio Seam behavior, but a valid preserved or directly connected Driving Audio source takes precedence and is returned instead. The output retains the selected audio source's sample rate and is adjusted for the review or exact-duration plan when required. Use this output for video encoding, playback, or downstream audio processing.

report

Returns a text summary of the assembly as a STRING. It can describe the selected processing behavior, audio source, seam handling, memory or buffer status, and other completion details. With diagnostics set to Detailed Report, it can also include projected assembly buffer information and the statement that V3.5 Exact Duration image writing is direct with no additional final image copy. Review this output when checking whether the intended audio source was selected or when diagnosing memory, duration, or seam behavior.

H3 Continuum Finalize Usage Tips:

  • Keep buffer_backend set to Auto for general use. It is designed to keep outputs up to 4 GiB in RAM only when a safety reserve is available and to use disk-backed storage when RAM would be unsafe.
  • Use RAM when you have confirmed that enough physical memory is available and you need the quickest memory-based assembly. Large image sequences can require substantial memory, so avoid forcing this mode on constrained systems.
  • Use Disk-backed for large outputs or when your system has limited free RAM. Ensure the drive used for temporary or mapped data has enough free space and good performance.
  • Enable exact_total_duration when the final video must match the assembly plan precisely or remain synchronized with an external timeline. Leave it disabled when preserving the natural assembled length is more important than exact plan duration.
  • Choose Auto for audio_seam when generated audio chunks meet at boundaries that may produce clicks, gaps, or abrupt changes. Choose Off when using clean source audio or when a connected Driving Audio source should remain untouched.
  • Use Analyze Only for video_seam to inspect potential visual boundary problems without applying corrections. After reviewing the report, select Auto or Auto 2 when correction is needed.
  • Connect driving_audio when the original recording must be preserved instead of using generated chunk audio. Check the report to confirm whether the selected source was direct input or a preserved source from the assembly plan.
  • Use Detailed Report while troubleshooting duration, seam, or memory issues. After the workflow is stable, Basic or Off can produce shorter reports.
  • Keep the image chunks, audio chunks, and assembly plan from the same Continuum generation or review workflow so their ordering, timing, and target frame information remain consistent.

H3 Continuum Finalize Common Errors and Solutions:

assembly_plan must contain exactly one value

  • Explanation: The node received no single assembly plan or received multiple plan values. The assembler requires one plan so it can determine chunk order, timing, and target duration.
  • Solution: Connect one valid H3_CONTINUUM_ASSEMBLY_PLAN value and remove extra values from the connection or list.

exact_total_duration must contain exactly one value

  • Explanation: The Boolean duration control was provided as multiple values instead of one value.
  • Solution: Connect or provide exactly one Boolean value, such as enabled or disabled, and ensure a list or batch is not being passed into this control.

audio_seam must contain exactly one value

  • Explanation: The Audio Seam selector received more than one value, so the node cannot determine whether audio seam processing should be enabled.
  • Solution: Set the control to one option only: Off or Auto.

video_seam must contain exactly one value

  • Explanation: The Video Seam selector received multiple values instead of one processing mode.
  • Solution: Select one option: Auto, Auto 2, Analyze Only, or Off.

buffer_backend must contain exactly one value

  • Explanation: The Buffer Backend selector received multiple values, preventing the node from choosing a single image-storage strategy.
  • Solution: Select one option: Auto, RAM, or Disk-backed.

diagnostics must contain exactly one value

  • Explanation: The diagnostics selector was passed more than one value, so the node cannot choose a reporting level.
  • Solution: Select one diagnostics mode: Basic, Detailed Report, or Off.

Projected assembly buffers [V3.5]: unavailable

  • Explanation: Detailed diagnostics could not calculate the projected image and audio buffer requirements. The report includes the specific exception type and message after this text. This diagnostic failure concerns the projection report and may not necessarily mean that the main assembly failed.
  • Solution: Review the exception details in the report, confirm that the image chunks, audio chunks, and assembly plan are valid and compatible, and retry with diagnostics set to Basic if you only need the assembled result.

Driving Audio is missing or invalid

  • Explanation: The optional audio value did not contain a valid waveform tensor and sample rate, so it could not replace the generated audio. The node then continues with generated or plan-provided audio when available.
  • Solution: Connect a valid AUDIO output containing waveform data and a sample rate. Confirm that the source is not an empty, unsupported, or incorrectly typed value.

Insufficient memory while using RAM

  • Explanation: The final image assembly was forced into RAM, but the system did not have enough available memory for the requested output.
  • Solution: Change buffer_backend to Auto or Disk-backed, reduce the output resolution or duration, close other memory-intensive applications, and retry.

Insufficient disk space while using Disk-backed

  • Explanation: The disk-backed image buffer could not be created or expanded because the selected drive lacks sufficient free space.
  • Solution: Free disk space, use a drive with more available capacity, or switch to Auto or RAM if adequate memory is available.

Video seam analysis or correction failure

  • Explanation: A selected Video Seam mode could not analyze or process one or more chunk boundaries. This can occur when chunks are incomplete, incompatible, or do not contain enough usable frame information for the selected operation.
  • Solution: Try video_seam set to Analyze Only to inspect the boundaries, then verify the source chunks and assembly plan. Use Off when the chunks are already continuous or use the other automatic mode if the issue is specific to one correction path.

Exact duration does not match the expected result

  • Explanation: The assembled sequence may not have the expected duration when the plan, input chunks, or audio timing do not agree. Exact duration enforcement can only use the target information available in the assembly plan.
  • Solution: Confirm that assembly_plan, images, and audio belong to the same generation or review pass, enable exact_total_duration, and check the returned report with Detailed Report selected.

Audio and image timing mismatch

  • Explanation: The audio and image chunks do not describe matching timing or duration, which can produce an output with drift, early audio ending, or extra audio.
  • Solution: Reconnect matching image and audio outputs from the same sampler process, verify the assembly plan, and enable exact_total_duration when the final output must follow the plan exactly.

Invalid or incomplete image/audio chunk input

  • Explanation: The node could not assemble the supplied chunks because one or more inputs are missing, empty, malformed, or incompatible with the expected IMAGE or AUDIO types.
  • Solution: Confirm that images receives decoded image chunks, audio receives matching audio chunks, and assembly_plan receives one valid Continuum plan. Re-run the upstream sampler or decoder if any chunk output is incomplete.

H3 Continuum Finalize Related Nodes

Go back to the extension to check out more related nodes.
H3 Continuum
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.

H3 Continuum Finalize