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 > OpenKatana > OpenKatana · Make captions

ComfyUI Node: OpenKatana · Make captions

Class Name

OpenKatana_make_captions

Category
OpenKatana/Tools
Author
rk0neki (Account age: 235days)
Extension
OpenKatana
Latest Updated
2026-11-05
Github Stars
0.05K

How to Install OpenKatana

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

OpenKatana · Make captions Description

Creates styled, timed captions from word-timing files with grouping and karaoke highlighting.

OpenKatana · Make captions:

OpenKatana_make_captions creates timed captions from an existing word-timing file for an OpenKatana project. It provides controls for caption appearance, timing, grouping, and karaoke-style highlighting, helping you turn a timed transcript into readable captions that suit your video’s visual style. The underlying caption generator works with word start and end times rather than transcribing audio itself, so preparing an accurate timed transcript is an important first step. You can adjust typography and control how much text is grouped together, making the tool useful for balancing readability with a compact presentation. The provided code does not establish whether captions are automatically added to a timeline or rendered into a finished video, so caption creation should not be treated as confirmation of either action.

OpenKatana · Make captions Input Parameters:

project_id

A required STRING that identifies the OpenKatana project for the caption-making request. Use the identifier of the project you intend to work with so that the request targets the correct project context. This is a project identifier, not the transcript filename or caption text. The supplied interface does not specify a default value, allowed options, length limits, or a particular identifier format; minimum and maximum values are not specified.

arguments

A required STRING containing the tool arguments, normally supplied as a JSON object encoded as text. This is where you configure caption generation; the settings inside it are not separate node sockets. The underlying API requires words, a path to the word-timing file, and family, the font-family setting. Its optional settings and API defaults are fps = 30, style = "clean", size = None when unspecified, height = 1080, offset = 0, highlight = "#ffd200", karaoke_mode = "word", max_chars = 32, max_words = 7, max_gap = 0.6, and max_duration = 3. The appearance settings control styling, font size, reference height, and highlighting; the timing and grouping settings provide controls for timing adjustment, text length, word count, gaps, and caption duration. The supplied excerpts do not show their full implementation, permitted ranges, or complete lists of supported styles and karaoke modes. These are underlying API defaults, not confirmed default text for the node’s arguments field, whose default is not shown. An example JSON string is {"words":"/path/to/words.json","family":"Arial","fps":30,"style":"clean","max_chars":32,"max_words":7}; replace the path and font family with suitable values for your environment. For JSON word-timing data, the provided loader accepts a list of entries shaped as [0.0, 0.4, "Hello"] or objects shaped as {"start":0.0,"end":0.4,"word":"Hello"}; object entries can also use t0, t1, and text as alternatives. Word start times must be nonnegative, end times must be at least their corresponding start times, and text must be a nonempty string.

after

An optional STRING accepted by the node’s execution method. In the supplied implementation, this value is not read, added to the tool arguments, or passed to the tool invocation, so it has no demonstrated effect on caption generation or the returned content. The Python execution method defaults it to None when omitted, but no STRING widget default, allowed options, or minimum and maximum lengths are shown. Do not use this field to set caption timing or delay; the underlying caption API exposes an offset setting through arguments for timing adjustment.

OpenKatana · Make captions Output Parameters:

result_json

A STRING containing the response from the OpenKatana tool call in JSON form. Use it to inspect the result of the caption-making request and determine what information the backend returned. The node passes the tool invocation through a response helper rather than exposing caption events, files, or cues as individual sockets. Although the underlying caption-generation excerpt mentions events and cues, the supplied code does not show the API response dictionary or the response helper’s serialization details, so specific JSON keys, generated file paths, success flags, and timeline changes cannot be documented reliably. Treat this output as a tool response to inspect, not as an image, video, or guaranteed subtitle-file path.

OpenKatana · Make captions Usage Tips:

  • Prepare word-level timing data before running the node. For example, [[0.0,0.4,"Hello"],[0.45,0.9,"world"]] follows the list-entry format accepted by the provided loader; plain transcript text without start and end times does not follow that format.

  • Begin with the API defaults, including style set to "clean" and karaoke_mode set to "word", then change one setting at a time. These values are explicitly shown in the API signature; other supported choices are not listed in the supplied context.

  • For short, compact captions, try lowering max_chars and max_words from their defaults of 32 and 7. Review the resulting grouping on representative dialogue rather than assuming a particular line layout, because the line-breaking implementation is not included.

  • Use max_gap and max_duration when adjusting grouping around pauses or long phrases. Their defaults are 0.6 and 3 respectively; test changes against both fast speech and slower passages to check readability.

  • Set fps to match your intended project frame rate and height to match the relevant output height. The API defaults are 30 and 1080, but the supplied context does not establish how changing these settings affects every aspect of caption layout or timing.

  • Set highlight to a color that remains visible against your footage; the default is the yellow hexadecimal color "#ffd200". Check bright and dark shots when judging contrast, and do not assume that a highlight color alone guarantees readable captions.

  • Keep timing adjustments in arguments through offset, not in after. Test an adjustment on a short section first, because the supplied excerpts do not document the offset’s direction or its interaction with project timing.

  • Inspect result_json after execution before connecting downstream steps that expect a caption path or other specific field. The response schema is not provided, so confirm the actual returned structure in your installation.

OpenKatana · Make captions Common Errors and Solutions:

unexpected inputs: ...

  • Explanation: The node explicitly raises a TypeError with this message prefix if it receives extra execution fields that are not declared in its typed input definitions. Caption settings such as fps and family are not listed as separate sockets in the authoritative interface.
  • Solution: Use only project_id, arguments, and the optional after input. Place caption-generation settings inside the JSON text supplied to arguments instead of adding extra node inputs.

filename: word N is not [t0, t1, word] or {start, end, word}

  • Explanation: The word-timing loader found an entry that was neither a list with at least three elements nor a dictionary. In the actual message, the filename and word index replace the placeholders; indexing starts at zero.
  • Solution: Replace the affected entry with a supported structure, such as [0.0,0.4,"Hello"] or {"start":0.0,"end":0.4,"word":"Hello"}. Check the other entries for plain strings, null values, or lists missing timing information.

filename: word N has no text

  • Explanation: The affected entry has no usable word text. The loader requires a string containing something other than whitespace; missing values, empty strings, numbers, and whitespace-only strings fail this check.
  • Solution: Supply meaningful text in the entry’s third list position or its word or text field. Remove empty entries that do not represent spoken words.

filename: no words

  • Explanation: The loader did not obtain any words. This can happen when the JSON list is empty or when the top-level JSON value is not a list; the shown code only iterates over a top-level list.
  • Solution: Supply a nonempty top-level list of valid word entries. For example, use [`{"start":0.0,"end":0.4,"word":"Hello"}`] rather than wrapping that list in an additional object.

Invalid word timing values

  • Explanation: This is a diagnostic description, not a confirmed exact error message. The provided validation requires numeric start and end times, a start time of at least zero, and an end time no earlier than the corresponding start. The validator’s exact error wording is not shown.
  • Solution: Check each entry for negative starts, reversed start and end times, missing times, or numbers written as quoted text. Use numeric values such as 0.4 rather than "0.4".

Invalid arguments JSON

  • Explanation: This is a diagnostic description; the argument parser’s exact error text is not included. A malformed JSON string cannot reliably supply the caption API with its settings.
  • Solution: Use a JSON object with double-quoted property names and string values, numeric values without quotes, and no trailing commas. Start with a small object containing words and family, then add optional settings after the basic request works.

Missing required caption arguments

  • Explanation: This is a diagnostic description rather than a confirmed error string. The underlying API signature requires words and family in addition to the project identifier, even though these settings do not appear as individual node inputs.
  • Solution: Include both words and family in arguments, and provide project_id through its required node input. Use words for the timing-file path, not for a block of untimed transcript text.

Word-timing file cannot be read

  • Explanation: This is a troubleshooting category, not an exact error message from the supplied excerpts. Caption generation depends on reading the file identified by words; an incorrect path, inaccessible file, or malformed file content can prevent that step.
  • Solution: Confirm that the file exists and is readable in the environment running OpenKatana, not only on your own computer. Check that it contains valid word-timing data in the supported structure, then retry with the corrected path.

OpenKatana · Make captions Related Nodes

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

OpenKatana · Make captions