docs: sync built-in nodes documentation (weekly update)

[lin-bot23]

Weekly sync of built-in node documentation from embedded-docs pipeline.

New node pages

• built-in-nodes/IdeogramPImage.mdx
• built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
• built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
• built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx

(en + zh/ja/ko translations)

Updated

• ~56 built-in node pages refreshed from embedded-docs en.md
• docs.json navigation updated
• Escaped <= in OpenAIGPTImageNodeV2 (MDX parser compatibility)

[coderabbitai]

📝 Walkthrough

Walkthrough

Changes

Built-in node documentation

Layer / File(s)	Summary
New node pages and navigation built-in-nodes/IdeogramPImage.mdx, built-in-nodes/MinimaxHailuo03*.mdx, */built-in-nodes/IdeogramPImage.mdx, */built-in-nodes/MinimaxHailuo03*.mdx, docs.json	Added IdeogramPImage and MiniMax Hailuo 03 documentation in four languages. Added localized navigation entries.
Media and video behavior built-in-nodes/BerniniConditioning.mdx, built-in-nodes/ByteDanceSeedAudio.mdx, built-in-nodes/GeminiVideoOmni.mdx, built-in-nodes/HeyGen*.mdx, built-in-nodes/Kling*.mdx, built-in-nodes/LumaRay32*.mdx, built-in-nodes/Runway*.mdx, built-in-nodes/Video*.mdx, built-in-nodes/SyncLipSyncNode.mdx, built-in-nodes/LoadVideo*Dataset*.mdx, */built-in-nodes/GeminiVideoOmni.mdx	Documented media limits, input requirements, video generation modes, dataset behavior, temporal sampling, and localized GeminiVideoOmni updates.
Geometry, model, and conditioning behavior built-in-nodes/DA3*.mdx, built-in-nodes/Load3DAdvanced.mdx, built-in-nodes/LoadDA3Model.mdx, built-in-nodes/Save*.mdx, built-in-nodes/TextEncode*.mdx, built-in-nodes/SCAIL2ColoredMask.mdx, built-in-nodes/TripoImportModelNode.mdx, built-in-nodes/WanUni3CControlnetApply.mdx	Added documentation for geometry filtering, model precision, 3D loading and saving, conditioning inputs, mask behavior, import constraints, and model compatibility.
General input and output clarifications built-in-nodes/BuildJsonPromptIdeogram.mdx, built-in-nodes/ConvertDictionaryToString.mdx, built-in-nodes/DA3Inference.mdx, built-in-nodes/OpenAIGPTImageNodeV2.mdx, built-in-nodes/SeedNode.mdx, built-in-nodes/SeedVR2Temporal*.mdx, built-in-nodes/ShuffleVideo*.mdx, built-in-nodes/SaveText.mdx, built-in-nodes/TextOverlay.mdx	Clarified defaults, prompt requirements, seed behavior, latent validation, shuffling, JSON formatting, overlay behavior, and Markdown formatting.
Node-specific constraints and lifecycle text built-in-nodes/AnimaLLLiteApply.mdx, built-in-nodes/Runway*.mdx, zh/built-in-nodes/Runway*.mdx	Updated experimental-status notes, API constraints, node descriptions, and obsolete lifecycle warnings.

Possibly related PRs

• Comfy-Org/docs#1338: Adds or updates localized documentation for the same Ideogram P Image and MiniMax H3 nodes.

🚥 Pre-merge checks | ✅ 2

✅ Passed checks (2 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

✨ Simplify code

• Create PR with simplified code

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 18

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@built-in-nodes/BuildJsonPromptIdeogram.mdx`:
- Line 17: Update the `background` row description in the node documentation to
replace the dash-joined clauses with two direct sentences: state that providing
a value is strongly recommended, then explain that an empty string still runs
but may produce unpredictable results.
- Around line 17-23: Update the descriptions for background, aesthetics,
lighting, and medium in the parameter table to say “Recommended for reliable
generation” instead of “Mandatory,” preserving their current optional status and
empty-string defaults.

In `@built-in-nodes/HeyGenVideoTranslateNode.mdx`:
- Line 17: Align the required metadata with each node’s actual defaulting
behavior: update built-in-nodes/HeyGenVideoTranslateNode.mdx line 17 and
built-in-nodes/LumaRay32ExtendVideoNode.mdx line 19 to mark the dropdown
optional if the corresponding node supplies its documented default when omitted;
otherwise remove the default or document why the input must remain required.

In `@built-in-nodes/IdeogramPImage.mdx`:
- Around line 15-20: Update the English descriptions in the prompt and seed rows
of the table to remove em dashes, replacing them with sentence breaks, colons,
commas, or parentheses while preserving the existing meaning.

In `@built-in-nodes/KlingSingleImageVideoEffectNode.mdx`:
- Line 9: Replace the introductory description for the Kling Single Image Video
Effect Node with a technical reference statement: it generates a video from one
reference image using the selected effect scene, model, and duration. Remove
promotional phrases such as “transform static images into dynamic video content”
and “desired visual outcome.”

In `@built-in-nodes/KlingVirtualTryOnNode.mdx`:
- Line 9: Update the description text for Kling Virtual Try On Node to replace
“cloth image” with “clothing image” or “garment image,” and state clearly that
the node applies the clothing to the human image. Keep the wording short,
direct, and technical.

In `@built-in-nodes/LoadVideoTextDataSetFromFolder.mdx`:
- Around line 23-24: Update the `videos` and `texts` table descriptions in
`LoadVideoTextDataSetFromFolder` to describe entries after numeric-prefix repeat
expansion rather than one entry per physical file. State that each video
reference is repeated as configured, and that `texts` contains the corresponding
repeated, index-aligned caption entries, including empty captions when no
matching `.txt` file exists.

In `@built-in-nodes/LumaRay32KeyframeNode.mdx`:
- Line 9: Remove the opening and closing Markdown code fences surrounding the
entire LumaRay32KeyframeNode documentation, leaving the headings, Inputs/Outputs
tables, and descriptions as normal Markdown content.

In `@built-in-nodes/LumaRay32KeyframesToVideoNode.mdx`:
- Line 15: Align the `prompt` schema definition with runtime validation: either
replace its empty-string default with a valid non-empty prompt or update the
declared length range to allow 0–6000 characters, ensuring the documented
default is accepted during execution.

In `@built-in-nodes/RunwayFirstLastFrameNode.mdx`:
- Line 9: Update the RunwayFirstLastFrameNode documentation to add an EOL status
section before the current supported-node description. State that the node is
retired and direct users to Gen4 Turbo or another supported replacement,
matching the EOL guidance in the related Runway video-generation documentation.

In `@built-in-nodes/SeedVR2TemporalChunk.mdx`:
- Around line 23-25: Update the frames_per_chunk range in the parameter table to
align with the 4n+1 constraint by using 1 to 16381, the largest valid value in
the documented range. Keep the surrounding description and validation behavior
unchanged.

In `@built-in-nodes/SyncLipSyncNode.mdx`:
- Around line 27-28: The speaker_y documentation currently implies the full
0–4096 range is valid for the vertical coordinate without explaining the
distinction from speaker_x. Update the speaker_y entry in SyncLipSyncNode to
document the correct 2160 height limit for 4K video, or explicitly note that
4096 is only a shared API maximum and not a valid vertical video coordinate.

In `@built-in-nodes/TextEncodeBooguEdit.mdx`:
- Line 17: Update the negative_prompt description in TextEncodeBooguEdit
documentation so the second sentence is complete and directly addresses the
user, using wording equivalent to “You can leave it empty to drop negative
conditioning.”

In `@built-in-nodes/TextEncodeMageFlowEdit.mdx`:
- Line 31: Update the overview in TextEncodeMageFlowEdit to explicitly
distinguish the two reference-image resize paths: resize images with the longest
edge capped at 384 pixels for text conditioning, while resizing them to the full
target output resolution for the VAE encoding branch. Replace the conflicting
statement that all references use the target resolution, preserving the rest of
the preprocessing description.

In `@built-in-nodes/TripoImportModelNode.mdx`:
- Around line 15-17: Remove the contradictory “Any 3D format” entry from the
`model_3d` format list in `TripoImportModelNode.mdx`, leaving only GLB, FBX,
OBJ, and STL to match the supported-format note.

In `@ko/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx`:
- Around line 21-22: Update the reference_videos and reference_audios
descriptions in the parameter table to state that their combined duration is at
most 15 seconds, not exactly 15 seconds. Replace the exact-duration wording with
equivalent “within 15 seconds” wording while preserving the existing per-item
limits and other guidance.

In `@zh/built-in-nodes/IdeogramPImage.mdx`:
- Line 9: Update the description in Ideogram P-Image to replace “强大的排版” with the
noun phrase “强大的排版能力”, leaving the rest of the description unchanged.

In `@zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx`:
- Around line 12-13: Translate the headers of both input and output parameter
tables in MinimaxHailuo03FirstLastFrameNode into consistent Chinese labels,
including Parameter, Description, Data Type, Required, and Range. Update both
table header rows while preserving their existing structure and column order.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: d48a5161-036b-48c5-b51b-207f7ad2fa27

📥 Commits

Reviewing files that changed from the base of the PR and between 45380d7 and c9ef422.

📒 Files selected for processing (76)

• built-in-nodes/AnimaLLLiteApply.mdx
• built-in-nodes/BerniniConditioning.mdx
• built-in-nodes/BuildJsonPromptIdeogram.mdx
• built-in-nodes/ByteDanceSeedAudio.mdx
• built-in-nodes/ConvertDictionaryToString.mdx
• built-in-nodes/DA3GeometryToMesh.mdx
• built-in-nodes/DA3GeometryToPointCloud.mdx
• built-in-nodes/DA3Inference.mdx
• built-in-nodes/DA3Render.mdx
• built-in-nodes/GeminiVideoOmni.mdx
• built-in-nodes/HeyGenAvatarVideoNode.mdx
• built-in-nodes/HeyGenCreateAvatarNode.mdx
• built-in-nodes/HeyGenTalkingPhotoNode.mdx
• built-in-nodes/HeyGenTextToSpeechNode.mdx
• built-in-nodes/HeyGenVideoTranslateNode.mdx
• built-in-nodes/IdeogramPImage.mdx
• built-in-nodes/KlingDualCharacterVideoEffectNode.mdx
• built-in-nodes/KlingImage2VideoNode.mdx
• built-in-nodes/KlingSingleImageVideoEffectNode.mdx
• built-in-nodes/KlingStartEndFrameNode.mdx
• built-in-nodes/KlingVirtualTryOnNode.mdx
• built-in-nodes/Load3DAdvanced.mdx
• built-in-nodes/LoadDA3Model.mdx
• built-in-nodes/LoadVideoDataSetFromFolder.mdx
• built-in-nodes/LoadVideoTextDataSetFromFolder.mdx
• built-in-nodes/LumaRay32ExtendVideoNode.mdx
• built-in-nodes/LumaRay32KeyframeNode.mdx
• built-in-nodes/LumaRay32KeyframesToVideoNode.mdx
• built-in-nodes/LumaRay32TextToVideoNode.mdx
• built-in-nodes/LumaRay32VideoEditNode.mdx
• built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
• built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
• built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
• built-in-nodes/OpenAIGPTImageNodeV2.mdx
• built-in-nodes/RunwayAleph2KeyframeNode.mdx
• built-in-nodes/RunwayAleph2PromptImageNode.mdx
• built-in-nodes/RunwayFirstLastFrameNode.mdx
• built-in-nodes/RunwayImageToVideoNodeGen3a.mdx
• built-in-nodes/SCAIL2ColoredMask.mdx
• built-in-nodes/Save3DAdvanced.mdx
• built-in-nodes/SaveGaussianSplat.mdx
• built-in-nodes/SavePointCloud.mdx
• built-in-nodes/SaveText.mdx
• built-in-nodes/SeedNode.mdx
• built-in-nodes/SeedVR2TemporalChunk.mdx
• built-in-nodes/SeedVR2TemporalMerge.mdx
• built-in-nodes/ShuffleVideoDataset.mdx
• built-in-nodes/ShuffleVideoTextDataset.mdx
• built-in-nodes/SyncLipSyncNode.mdx
• built-in-nodes/TextEncodeBooguEdit.mdx
• built-in-nodes/TextEncodeJoyImageEdit.mdx
• built-in-nodes/TextEncodeMageFlowEdit.mdx
• built-in-nodes/TextOverlay.mdx
• built-in-nodes/TripoImportModelNode.mdx
• built-in-nodes/VideoFrameSample.mdx
• built-in-nodes/VideoRandomTemporalCrop.mdx
• built-in-nodes/VideoTemporalCrop.mdx
• built-in-nodes/WanUni3CControlnetApply.mdx
• docs.json
• ja/built-in-nodes/GeminiVideoOmni.mdx
• ja/built-in-nodes/IdeogramPImage.mdx
• ja/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
• ja/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
• ja/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
• ko/built-in-nodes/GeminiVideoOmni.mdx
• ko/built-in-nodes/IdeogramPImage.mdx
• ko/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
• ko/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
• ko/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
• zh/built-in-nodes/GeminiVideoOmni.mdx
• zh/built-in-nodes/IdeogramPImage.mdx
• zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx
• zh/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx
• zh/built-in-nodes/MinimaxHailuo03TextToVideoNode.mdx
• zh/built-in-nodes/RunwayFirstLastFrameNode.mdx
• zh/built-in-nodes/RunwayImageToVideoNodeGen3a.mdx

💤 Files with no reviewable changes (2)

• built-in-nodes/ConvertDictionaryToString.mdx
• built-in-nodes/RunwayImageToVideoNodeGen3a.mdx

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Split the dash-joined clauses into two sentences.

The background description joins two independent clauses with a dash. Write: Providing a value is strongly recommended. The node will run with an empty string, but results may be unpredictable. No dash, less clash.

As per coding guidelines: In English MDX, prefer short, direct sentences over stacked clauses joined by dashes.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/BuildJsonPromptIdeogram.mdx` at line 17, Update the
`background` row description in the node documentation to replace the
dash-joined clauses with two direct sentences: state that providing a value is
strongly recommended, then explain that an empty string still runs but may
produce unpredictable results.

Source: Coding guidelines

[coderabbitai]

🗄️ Data Integrity & Integration | 🟡 Minor | ⚡ Quick win

Resolve the contradictory requiredness guidance.

The table marks background, aesthetics, lighting, and medium as Required: No, but their descriptions call them “Mandatory”. The note also says empty strings are allowed. Use one term consistently. If empty values are valid, use “Recommended for reliable generation”. If the node enforces these values, mark them as required and document the validation.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/BuildJsonPromptIdeogram.mdx` around lines 17 - 23, Update the
descriptions for background, aesthetics, lighting, and medium in the parameter
table to say “Recommended for reliable generation” instead of “Mandatory,”
preserving their current optional status and empty-string defaults.

[coderabbitai]

🗄️ Data Integrity & Integration | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
set -euo pipefail

rg -n -C 6 'HeyGenVideoTranslateNode|LumaRay32ExtendVideoNode|mode|resolution' .

Repository: Comfy-Org/docs

Length of output: 50371

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

echo "== matching markdown files =="
git ls-files | rg -n '(^|/)HeyGenVideoTranslateNode\.mdx$|(^|/)LumaRay32ExtendVideoNode\.mdx$|HeyGenVideoTranslateNode|LumaRay32ExtendVideoNode'

echo
echo "== english table lines =="
for f in Built-in-nodes built-in-nodes; do
  [ -d "$f" ] && find "$f" -type f \( -path '*HeyGenVideoTranslateNode.mdx' -o -path '*LumaRay32ExtendVideoNode.mdx' \) | while read -r file; do
    echo "--- $file"
    rg -n -C 2 'mode|resolution|speed|precision|720p|4k|1080p' "$file"
  done
done

echo
echo "== repo-wide node definitions and docs only =="
rg -n --glob '!ko/**' --glob '!**/node_type_schema.json' \
  -C 5 'class HeyGenVideoTranslateNode|class LumaRay32ExtendVideoNode|function HeyGenVideoTranslateNode|function LumaRay32ExtendVideoNode|HeyGenVideoTranslateNode|LumaRay32ExtendVideoNode|max=|required|default|speed|precision|resolution' .

echo
echo "== parse relevant markdown input rows =="
python3 - <<'PY'
from pathlib import Path
matches = [p for p in Path('.').rglob('*.mdx') if p.parts[0] not in {'ko'} and ('HeyGenVideoTranslateNode' in str(p) or 'LumaRay32ExtendVideoNode' in str(p))]
for p in matches:
    lines = p.read_text(errors='replace').splitlines()
    print(f"\n--- {p} ---")
    for i,l in enumerate(lines,1):
        if '`mode`' in l or '`resolution`' in l or l.startswith('|---'):
            print(f"{i}: {l}")
PY

Repository: Comfy-Org/docs

Length of output: 50373

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

echo "== matching English/Markdown docs only =="
git ls-files | rg -n '^(built-in-nodes/(HeyGenVideoTranslateNode\.mdx|LumaRay32ExtendVideoNode\.mdx))$|HeyGenVideoTranslateNode|LumaRay32ExtendVideoNode' | head -80

echo
echo "== English table rows for the two pages =="
for f in built-in-nodes/HeyGenVideoTranslateNode.mdx built-in-nodes/LumaRay32ExtendVideoNode.mdx; do
  [ -f "$f" ] && {
    echo "--- $f"
    sed -n '1,60p' "$f" | cat -n
  }
done

echo
echo "== node source/docs references (English docs + tracked code-like docs) =="
rg -n --glob '!ko/**' --glob '!zh/**' --glob '!**/comfyui_embedded_docs/**' \
  -C 8 'HeyGenVideoTranslateNode|LumaRay32ExtendVideoNode' .

Repository: Comfy-Org/docs

Length of output: 24862

---

Align required/default metadata with the node behavior.

Both HeyGenVideoTranslateNode.mdx#L17 and LumaRay32ExtendVideoNode.mdx#L19 show a default while marking the input required. If the node actually supplies the default when the input is omitted, mark the field optional; otherwise document why omission is not allowed. A required dropdown with a default is still a required requirement, but this row suggests otherwise.

📍 Affects 2 files

• built-in-nodes/HeyGenVideoTranslateNode.mdx#L17-L17 (this comment)
• built-in-nodes/LumaRay32ExtendVideoNode.mdx#L19-L19

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/HeyGenVideoTranslateNode.mdx` at line 17, Align the required
metadata with each node’s actual defaulting behavior: update
built-in-nodes/HeyGenVideoTranslateNode.mdx line 17 and
built-in-nodes/LumaRay32ExtendVideoNode.mdx line 19 to mark the dropdown
optional if the corresponding node supplies its documented default when omitted;
otherwise remove the default or document why the input must remain required.

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Replace the em dashes in the English table text.

Lines 15 and 20 use em dashes. Replace them with sentence breaks or colons. One dash too many can make the style checker dash.

Suggested wording

-| `prompt` | ... bounding-box layout) — set prompt_upsampling to OFF ...
+| `prompt` | ... bounding-box layout). Set prompt_upsampling to OFF ...

-| `seed` | ... prompt rewrite varies per run — reproduce a result ...
+| `seed` | ... prompt rewrite varies per run. Reproduce a result ...

As per coding guidelines: In English documentation, avoid em dashes. Use periods, commas, colons, parentheses, or a second sentence.

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	| `prompt` | Text prompt. Also accepts an Ideogram 4.0 structured JSON caption (exact colors as #RRGGBB hexes, exact text strings, bounding-box layout) — set prompt_upsampling to OFF to use it verbatim. Must not be empty. (default: "") | STRING | Yes | Any text |
	| `quality` | Speed/price/quality tier. MEDIUM is the everyday default; HIGH for complex prompts, fine detail and difficult text; VERY_LOW/LOW for drafts at scale. Difficult text renders poorly below MEDIUM. (default: "MEDIUM") | STRING | Yes | "VERY_LOW"<br />"LOW"<br />"MEDIUM"<br />"HIGH" |
	| `resolution` | Output size class (exact pixels follow the aspect ratio, e.g. 16:9 gives 1280x720 at 1K and 2560x1440 at 2K). Prefer HIGH + 2K for crisp typography. (default: "1K") | STRING | Yes | "1K"<br />"2K" |
	| `aspect_ratio` | The aspect ratio for image generation. (default: "1:1") | STRING | Yes | "1:3"<br />"3:1"<br />"1:2"<br />"2:1"<br />"9:16"<br />"16:9"<br />"10:16"<br />"16:10"<br />"2:3"<br />"3:2"<br />"3:4"<br />"4:3"<br />"4:5"<br />"5:4"<br />"1:1" |
	| `prompt_upsampling` | Expands short prompts into a detailed structured caption before generation (the rewritten prompt is returned as final_prompt). Set OFF when supplying your own JSON caption or exact wording. (default: "AUTO") | STRING | Yes | "AUTO"<br />"ON"<br />"OFF" |
	| `seed` | Seed for reproducible generation. With prompt_upsampling OFF, the same seed and settings return the same image; with ON/AUTO the prompt rewrite varies per run — reproduce a result by reusing its final_prompt output with prompt_upsampling OFF and the same seed. (default: 42) | INT | No | 0 to 2147483647 |
	| `prompt` | Text prompt. Also accepts an Ideogram 4.0 structured JSON caption (exact colors as `#RRGGBB` hexes, exact text strings, bounding-box layout). Set prompt_upsampling to OFF to use it verbatim. Must not be empty. (default: "") | STRING | Yes | Any text |
	| `quality` | Speed/price/quality tier. MEDIUM is the everyday default; HIGH for complex prompts, fine detail and difficult text; VERY_LOW/LOW for drafts at scale. Difficult text renders poorly below MEDIUM. (default: "MEDIUM") | STRING | Yes | "VERY_LOW"<br />"LOW"<br />"MEDIUM"<br />"HIGH" |
	| `resolution` | Output size class (exact pixels follow the aspect ratio, e.g. 16:9 gives 1280x720 at 1K and 2560x1440 at 2K). Prefer HIGH + 2K for crisp typography. (default: "1K") | STRING | Yes | "1K"<br />"2K" |
	| `aspect_ratio` | The aspect ratio for image generation. (default: "1:1") | STRING | Yes | "1:3"<br />"3:1"<br />"1:2"<br />"2:1"<br />"9:16"<br />"16:9"<br />"10:16"<br />"16:10"<br />"2:3"<br />"3:2"<br />"3:4"<br />"4:3"<br />"4:5"<br />"5:4"<br />"1:1" |
	| `prompt_upsampling` | Expands short prompts into a detailed structured caption before generation (the rewritten prompt is returned as final_prompt). Set OFF when supplying your own JSON caption or exact wording. (default: "AUTO") | STRING | Yes | "AUTO"<br />"ON"<br />"OFF" |
	| `seed` | Seed for reproducible generation. With prompt_upsampling OFF, the same seed and settings return the same image; with ON/AUTO the prompt rewrite varies per run. Reproduce a result by reusing its final_prompt output with prompt_upsampling OFF and the same seed. (default: 42) | INT | No | 0 to 2147483647 |

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/IdeogramPImage.mdx` around lines 15 - 20, Update the English
descriptions in the prompt and seed rows of the table to remove em dashes,
replacing them with sentence breaks, colons, commas, or parentheses while
preserving the existing meaning.

Source: Coding guidelines

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Use a technical description instead of promotional wording.

“Transform static images into dynamic video content” and “desired visual outcome” do not describe specific node behavior. State that the node generates a video from one reference image with the selected effect scene, model, and duration.

Proposed wording

-The Kling Single Image Video Effect Node creates videos with different special effects based on a single reference image. It applies various visual effects and scenes to transform static images into dynamic video content. The node supports different effect scenes, model options, and video durations to achieve the desired visual outcome.
+This node generates a video from one reference image using the selected effect scene, model, and duration.

As per coding guidelines, English MDX should use a technical-reference tone rather than a marketing tone.

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	The Kling Single Image Video Effect Node creates videos with different special effects based on a single reference image. It applies various visual effects and scenes to transform static images into dynamic video content. The node supports different effect scenes, model options, and video durations to achieve the desired visual outcome.
	This node generates a video from one reference image using the selected effect scene, model, and duration.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/KlingSingleImageVideoEffectNode.mdx` at line 9, Replace the
introductory description for the Kling Single Image Video Effect Node with a
technical reference statement: it generates a video from one reference image
using the selected effect scene, model, and duration. Remove promotional phrases
such as “transform static images into dynamic video content” and “desired visual
outcome.”

Source: Coding guidelines

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

Clarify the two reference-image resize paths.

Line 11 says that all reference images are resized to the target output resolution. Line 31 adds a separate 384-pixel path for text conditioning and a full-resolution path for VAE encoding. Update the overview so it describes both paths. Otherwise, users can expect the wrong conditioning input size.

Proposed wording

- It resizes all reference images to the target output resolution, encodes them into latent space if a VAE is provided, ...
+ It resizes reference images to a maximum 384-pixel longest edge for text conditioning and to the target output resolution for VAE encoding, if a VAE is provided, ...

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/TextEncodeMageFlowEdit.mdx` at line 31, Update the overview in
TextEncodeMageFlowEdit to explicitly distinguish the two reference-image resize
paths: resize images with the longest edge capped at 384 pixels for text
conditioning, while resizing them to the full target output resolution for the
VAE encoding branch. Replace the conflicting statement that all references use
the target resolution, preserving the rest of the preprocessing description.

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

Remove the contradictory Any 3D format entry.

Line 15 advertises Any 3D format, but Line 17 says only GLB, FBX, OBJ, and STL are supported. Remove the generic entry, or update the note only if the implementation accepts more formats. Keep the format list exact: supported in, unsupported out.

Proposed fix

-| `model_3d` | 3D model to import (GLB / FBX / OBJ / STL, up to 150 MB). OBJ and STL files carry no embedded textures. | FILE3D | Yes | GLB<br />FBX<br />OBJ<br />STL<br />Any 3D format |
+| `model_3d` | 3D model to import (GLB / FBX / OBJ / STL, up to 150 MB). OBJ and STL files carry no embedded textures. | FILE3D | Yes | GLB<br />FBX<br />OBJ<br />STL |

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	| `model_3d` | 3D model to import (GLB / FBX / OBJ / STL, up to 150 MB). OBJ and STL files carry no embedded textures. | FILE3D | Yes | GLB<br />FBX<br />OBJ<br />STL<br />Any 3D format |
	**Note:** GLB format is recommended because textures are preserved only when embedded directly in the file. OBJ and STL files do not support embedded textures. GLTF (.gltf) format is not supported because it references external files; use a single-file GLB instead. The file size must not exceed 150 MB.
	**Note:** Only GLB, FBX, OBJ, and STL formats are supported. GLTF (.gltf) cannot be imported because it references external files; use a single-file GLB instead. The model file must be 150 MB or smaller. GLB is recommended because textures survive import only when embedded in the file. OBJ and STL files do not carry embedded textures. Texturing an imported model requires a texture prompt.
	| `model_3d` | 3D model to import (GLB / FBX / OBJ / STL, up to 150 MB). OBJ and STL files carry no embedded textures. | FILE3D | Yes | GLB<br />FBX<br />OBJ<br />STL |
	**Note:** Only GLB, FBX, OBJ, and STL formats are supported. GLTF (.gltf) cannot be imported because it references external files; use a single-file GLB instead. The model file must be 150 MB or smaller. GLB is recommended because textures survive import only when embedded in the file. OBJ and STL files do not carry embedded textures. Texturing an imported model requires a texture prompt.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@built-in-nodes/TripoImportModelNode.mdx` around lines 15 - 17, Remove the
contradictory “Any 3D format” entry from the `model_3d` format list in
`TripoImportModelNode.mdx`, leaving only GLB, FBX, OBJ, and STL to match the
supported-format note.

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

Document the 15-second limit as a maximum.

Lines 21-22 say 총 15초입니다, which states that the combined duration is exactly 15 seconds. Lines 30-31 correctly state that the total must not exceed 15 seconds. Change both table entries to 총 15초 이내입니다 or equivalent. Maximum means maximum, not exact.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@ko/built-in-nodes/MinimaxHailuo03ReferenceNode.mdx` around lines 21 - 22,
Update the reference_videos and reference_audios descriptions in the parameter
table to state that their combined duration is at most 15 seconds, not exactly
15 seconds. Replace the exact-duration wording with equivalent “within 15
seconds” wording while preserving the existing per-item limits and other
guidance.

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Use a noun phrase for the typography capability.

强大的排版 is awkward in this sentence. Use 强大的排版能力.

Suggested wording

-该模型以强大的排版和逼真效果著称。
+该模型以强大的排版能力和逼真效果著称。

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	Ideogram P-Image 使用 Ideogram 的高速文本到图像模型，根据文本提示生成图像，该模型以强大的排版和逼真效果著称。它还支持 Ideogram 4.0 结构化 JSON 字幕，可精确控制文本字符串、颜色和布局。此节点返回生成的图像以及实际用于生成图像的最后提示词。
	Ideogram P-Image 使用 Ideogram 的高速文本到图像模型，根据文本提示生成图像，该模型以强大的排版能力和逼真效果著称。它还支持 Ideogram 4.0 结构化 JSON 字幕，可精确控制文本字符串、颜色和布局。此节点返回生成的图像以及实际用于生成图像的最后提示词。

🧰 Tools

🪛 LanguageTool

[uncategorized] ~9-~9: 动词的修饰一般为‘形容词（副词）+地+动词’。您的意思是否是：强大"地"排版
Context: ...使用 Ideogram 的高速文本到图像模型，根据文本提示生成图像，该模型以强大的排版和逼真效果著称。它还支持 Ideogram 4.0 结构化 JSON 字幕，...

(wb4)

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@zh/built-in-nodes/IdeogramPImage.mdx` at line 9, Update the description in
Ideogram P-Image to replace “强大的排版” with the noun phrase “强大的排版能力”, leaving the
rest of the description unchanged.

Source: Linters/SAST tools

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

Translate both table header rows.

The input and output tables retain English headers. Use consistent Chinese labels.

Suggested wording

-| Parameter | Description | Data Type | Required | Range |
+| 参数 | 描述 | 数据类型 | 必需 | 范围 |

-| Output Name | Description | Data Type |
+| 输出名称 | 描述 | 数据类型 |

Also applies to: 27-28

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@zh/built-in-nodes/MinimaxHailuo03FirstLastFrameNode.mdx` around lines 12 -
13, Translate the headers of both input and output parameter tables in
MinimaxHailuo03FirstLastFrameNode into consistent Chinese labels, including
Parameter, Description, Data Type, Required, and Range. Update both table header
rows while preserving their existing structure and column order.