n8n-nodes-kie-pro

25 n8n community nodes for Kie.ai — the unified AI API covering image generation, video generation, audio, music, chat, and upscaling. One API key. All models.

Includes Flux-2 Pro, Flux Kontext, Sora 2, Kling 3.0, Veo3, Wan 2.6, Hailuo 2.3, Runway Gen4, Google Imagen4, Nano Banana, Seedream 5.0, Seedance/Bytedance, Qwen Image 2.0, GPT-Image-1.5, Grok Imagine (incl. video + extend), Ideogram v3, Recraft, Topaz, ElevenLabs, Suno V5.5 (incl. Sounds + Mashup), ZImage, InfineTalk, Gemini (2.5/3/3.1), Claude (4 series), OpenAI Chat (GPT-5/Codex), Midjourney (Niji 7).

Installation

Via n8n GUI (Recommended)

1. Open your n8n instance → Settings → Community Nodes
2. Click Install
3. Enter: n8n-nodes-kie-pro
4. Click Install and restart n8n

Docker (CLI)

docker exec -u node n8n sh -c "
  mkdir -p /home/node/.n8n/nodes &&
  cd /home/node/.n8n/nodes &&
  npm install n8n-nodes-kie-pro
"
docker restart n8n

Bare Metal

cd ~/.n8n/nodes
npm install n8n-nodes-kie-pro

After restarting n8n, search for node names in the node picker (e.g. "Kling", "Suno", "Veo", "Flux").

API Key Setup

1. Visit kie.ai/api-key to generate your API key
2. In n8n → Credentials → New Credential → Kie API
3. Paste your API key and save

Available Nodes (25 Total)

🎬 Video Generation

Sora 2 Pro

Operations: Text-to-Video, Image-to-Video, Characters, Storyboard, Remove Watermark, Query Task Status

Models: sora-2-pro-text-to-video, sora-2-text-to-video, sora-2-characters-pro, sora-2-characters

Key parameters:

• prompt — text description of the video
• imageUrl — source image for I2V
• model — model variant
• seed — reproducibility seed (0 = random)
• replyUrl / replyRef — webhook callback URL and custom reference
• waitForCompletion — poll until done (default: true)

Kling

Operations: Text-to-Video, Image-to-Video, Video-to-Video (Motion Control), AI Avatar, Query Task Status

Models (T2V): kling-3.0/video ✨, kling-2.6/text-to-video, kling/v2-5-turbo-text-to-video-pro, kling/v2-1-master-text-to-video

Models (I2V): kling-2.6/image-to-video ✨, kling/v2-5-turbo-image-to-video-pro, kling/v2-1-master-image-to-video, kling/v2-1-pro, kling/v2-1-standard

Models (V2V): kling-3.0/motion-control, kling-2.6/motion-control

Key parameters:

• prompt, imageUrl
• model / modelI2V — model variant
• ratio — aspect ratio (16:9, 9:16, 1:1)
• duration — 5 or 10 seconds (Kling 3.0 supports 3–15s)
• generationMode — std or pro (Kling 3.0 only)
• enableSound — boolean, enables native audio (Kling 3.0 only)
• replyUrl / replyRef — webhook
• waitForCompletion

Seedance

Operations: Text-to-Video, Image-to-Video, Query Task Status

Models: bytedance/seedance-2, bytedance/seedance-2-fast, bytedance/seedance-1.5-pro, bytedance/v1-pro-text-to-video, bytedance/v1-lite-text-to-video, bytedance/v1-pro-image-to-video, bytedance/v1-lite-image-to-video, bytedance/v1-pro-fast-image-to-video

Key parameters:

• prompt, imageUrl
• model, ratio, duration (4/8/12/15s)
• seed, lastFrameUrl, resolution, generateAudio, returnLastFrame, webSearch
• referenceImageUrls, referenceVideoUrls, referenceAudioUrls
• replyUrl, replyRef, captchaToken
• waitForCompletion

Veo

Operations: Generate, Extend, Get 1080p Video, Get 4K Video, Query Task Status

Models: veo3 (Veo 3.1 Quality), veo3_fast (Veo 3.1 Fast), veo3_lite (Veo 3.1 Lite)

Key parameters:

• prompt, model — veo3, veo3_fast, or veo3_lite
• imageUrl — source image for I2V
• referenceUrls — multiple reference images
• endImageUrl — end frame
• ratio — 16:9, 9:16, or Auto
• generationType — auto, first/last frame, or reference-to-video
• duration — 5 or 8 seconds
• seed, watermark, replyUrl, replyRef
• waitForCompletion

Wan

Operations: Text-to-Video, Image-to-Video, Video-to-Video, Speech-to-Video, Animate, Image, Query Task Status

Models (T2V): wan/2-6-text-to-video, wan/2-5-text-to-video, wan/2-2-a14b-text-to-video-turbo

Models (I2V): wan/2-6-image-to-video, wan/2-6-flash-image-to-video ✨, wan/2-5-image-to-video, wan/2-2-a14b-image-to-video-turbo

Models (V2V): wan/2-6-video-to-video, wan/2-6-flash-video-to-video ✨, wan/2-2-animate-move, wan/2-2-animate-replace

Models (Image): wan/2-7-image, wan/2-7-image-pro

Key parameters:

• prompt, imageUrl, videoUrl
• model, ratio, duration
• seed, endImageUrl, replyUrl, replyRef, captchaToken
• imagePrompt, inputImageUrl, additionalInputImageUrls
• imageCount, enableSequential, imageResolution, thinkingMode, imageWatermark, imageSeed
• waitForCompletion

Hailuo

Operations: Text-to-Video, Image-to-Video, Query Task Status

Models (T2V): hailuo/02-text-to-video-pro, hailuo/02-text-to-video-standard

Models (I2V): hailuo/2-3-image-to-video-pro ✨, hailuo/2-3-image-to-video-standard ✨, hailuo/02-image-to-video-pro

Key parameters:

• prompt, imageUrl
• model, ratio, duration
• seed, endImageUrl, replyUrl, replyRef, captchaToken
• waitForCompletion

Runway

Operations: Generate, Extend, Aleph Generate, Query Task Status

Models: gen4_turbo (Gen4 Turbo), gen4 (Gen4)

Key parameters:

• prompt, imageUrl, endImageUrl
• model — gen4_turbo or gen4
• quality — 720p or 1080p (required by Runway API)
• ratio — aspect ratio
• duration — 5 or 10 seconds
• seed, replyUrl, replyRef
• waitForCompletion

GrokImagine

Operations: Text-to-Image, Image-to-Image, Text-to-Video, Image-to-Video, Upscale, Extend Video ✨, Query Task Status

Models: grok-imagine/text-to-image, grok-imagine/image-to-image, grok-imagine/text-to-video, grok-imagine/image-to-video, grok-imagine/upscale, grok-imagine/extend

Key parameters:

• prompt, imageUrl
• aspectRatio, videoAspectRatio, mode, duration, resolution, seed
• outputFormat — JPEG or PNG
• replyUrl, replyRef
• waitForCompletion

🖼️ Image Generation

Flux

Operations: Text-to-Image, Image-to-Image, Kontext Generate/Edit, Query Task Status

Models (T2I): flux-2/pro-text-to-image, flux-2/flex-text-to-image

Models (I2I): flux-2/pro-image-to-image, flux-2/flex-image-to-image

Models (Kontext): flux-kontext-pro, flux-kontext-max

Key parameters:

• prompt, inputImage
• model, aspectRatio, resolution — 1K or 2K (T2I/I2I)
• steps — inference steps (Kontext)
• strength — image strength for I2I (0–1)
• seed, outputFormat (JPEG/PNG/WebP)
• replyUrl, replyRef
• waitForCompletion

Seedream

Operations: Text-to-Image, Image Edit, Image-to-Image, Query Task Status

Models: seedream/5-lite-text-to-image, seedream/4.5-text-to-image, bytedance/seedream-v4-text-to-image, seedream/4.5-edit, bytedance/seedream-v4-edit, seedream/5-lite-image-to-image

Key parameters:

• prompt, imageUrl
• model, aspectRatio
• seed, outputFormat
• replyUrl, replyRef
• waitForCompletion

Google (Imagen4 + Nano Banana)

Operations: Generate, Edit, Image-to-Image, Query Task Status

Models: google/imagen4-ultra, google/imagen4, google/imagen4-fast, nano-banana-pro, nano-banana-2, google/nano-banana, google/nano-banana-edit

Key parameters:

• prompt, imageUrl
• model, aspectRatio
• seed, outputFormat, quality
• replyUrl, replyRef
• waitForCompletion

Ideogram

Operations: Generate (V3), Edit (V3) ✨, Remix (V3) ✨, Reframe (V3) ✨, Character, Character Edit, Character Remix, Query Task Status

Models: ideogram/v3-text-to-image, ideogram/v3-edit, ideogram/v3-remix, ideogram/v3-reframe, ideogram/character, ideogram/character-edit, ideogram/character-remix

Key parameters:

• prompt, imageUrl
• model, aspectRatio
• renderingSpeed — TURBO/DEFAULT/QUALITY
• seed, outputFormat
• referenceUrls, maskUrl
• replyUrl, replyRef
• waitForCompletion

Qwen

Operations: Text-to-Image, Image-to-Image, Image Edit, Query Task Status

Models: qwen2/text-to-image, qwen/text-to-image, qwen2/image-to-image ✨, qwen/image-to-image, qwen/image-edit, qwen2/image-edit

Key parameters:

• prompt, imageUrl
• model, aspectRatio
• seed, outputFormat
• replyUrl, replyRef
• waitForCompletion

GptImage15

Operations: Text-to-Image, Image-to-Image, Query Task Status

Models: gpt-image/1.5-text-to-image, gpt-image/1.5-image-to-image

Key parameters:

• prompt, imageUrl
• aspectRatio, quality (standard/hd)
• seed, outputFormat
• replyUrl, replyRef
• waitForCompletion

FourOImage

Operations: Generate, Query Task Status

Models: gpt-4o-image, gpt-4o-image-standard, gpt-4o-image-hd

Key parameters:

• prompt, model, aspectRatio
• seed, outputFormat
• replyUrl, replyRef
• waitForCompletion

ZImage

Operations: Generate, Query Task Status

Model: z-image

Key parameters:

• prompt, aspectRatio
• seed, outputFormat
• replyUrl, replyRef
• waitForCompletion

Recraft

Operations: Remove Background, Crisp Upscale, Query Task Status

Models: recraft/remove-background, recraft/crisp-upscale

Key parameters:

• imageUrl, seed, outputFormat
• replyUrl, replyRef
• waitForCompletion

Topaz

Operations: Image Upscale, Video Upscale, Query Task Status

Models: topaz/image-upscale, topaz/video-upscale

Key parameters:

• imageUrl / videoUrl, scale
• outputFormat
• replyUrl, replyRef
• waitForCompletion

🎵 Audio & Music

ElevenLabs

Operations: Text-to-Speech, Text-to-Dialogue V3 ✨, Speech-to-Text, Sound Effects, Audio Isolation, Query Task Status

Models: elevenlabs/text-to-speech-turbo-2-5, elevenlabs/text-to-speech-multilingual-v2, elevenlabs/text-to-dialogue-v3

Key parameters:

• text — text to speak
• ttsModel, voice (Rachel, Aria, Roger, Sarah, Laura, etc.)
• stability, similarityBoost
• dialogue — array of { voiceId, text } turns (Text-to-Dialogue)
• waitForCompletion

Suno

Operations: Generate Music, Extend Music, Upload & Cover Audio, Upload & Extend Audio, Add Instrumental, Add Vocals, Replace Music Section, Generate Lyrics, Boost Style, Convert to WAV, Separate Vocals, Generate MIDI, Create Music Video, Generate Sound ✨, Music Mashup ✨, Query Task Status

Models: V5_5 ✨, V5, V4_5PLUS, V4_5ALL, V4_5, V4, V3_5

Key parameters:

• prompt — style/lyric prompt
• modelVersion — V5, V4_5PLUS, V4_5ALL, V4_5, V4, or V3_5
• tags — style tags (e.g. "pop, female vocals")
• title, lyrics, style, instrumental
• sourceTaskId — for extend/boost/midi/video operations
• uploadUrl — audio URL for upload operations
• replyUrl, replyRef
• waitForCompletion

InfineTalk

Operations: Dub from Audio, Query Task Status

Model: infinetalk/from-audio

Key parameters:

• audioUrl — source audio file
• targetLanguage
• replyUrl, replyRef
• waitForCompletion

💬 Chat / LLM

Claude (Kie.ai) ✨

Operations: Chat Completion

Models: claude-opus-4-6 ✨, claude-sonnet-4-6, claude-haiku-4-5, claude-sonnet-4-5, claude-opus-4-5

Key parameters:

• model — Claude model to use
• message — user message
• systemPrompt — optional system instructions
• chatHistory — JSON array of {role, content} for multi-turn conversations
• imageUrl — optional image for multimodal inputs
• temperature, maxTokens
• Returns reply + updatedHistory for chaining

OpenAI Chat (Kie.ai) ✨

Operations: Chat Completion

Models: chat/gpt-5-2 ✨, market/codex/gpt-codex

Key parameters:

• model — GPT-5.2 or GPT Codex
• message, systemPrompt, chatHistory
• reasoningEffort — low/medium/high
• webSearch — enable Google Search grounding
• Returns reply + updatedHistory

Gemini (Kie.ai) ✨

Operations: Chat Completion

Models: gemini/gemini-3.1-pro ✨, gemini/gemini-3-pro, gemini/gemini-2.5-pro, gemini/gemini-2.5-flash

Key parameters:

• model, message, systemPrompt, chatHistory
• imageUrl — multimodal support
• googleSearch — enable real-time search grounding
• thinkingBudget — reasoning tokens
• Returns reply + updatedHistory

🎨 Image Generation (continued)

Midjourney (Kie.ai) ✨

Operations: Text-to-Image, Image-to-Image, Query Task Status

Models: Midjourney Standard, Niji 7 (anime)

Key parameters:

• prompt, imageUrl
• model — Standard or Niji 7
• mode — relaxed / fast / turbo
• aspectRatio, seed
• replyUrl, replyRef
• waitForCompletion

Features

• waitForCompletion on all create operations (on by default) — polls every 3s up to 5 min
• resultUrls surfaced at top level in every response for direct file access
• Webhook support via replyUrl + replyRef on all nodes
• seed parameter on all generation nodes for reproducible outputs
• Bearer token auth — single Kie.ai API key for all 22 nodes
• Exponential backoff on 429 rate limit errors (2s / 4s / 8s retries)
• continueOnFail support on all nodes

Usage Pattern

1. Drop a node into your workflow (e.g. Kling)
2. Select the operation (e.g. Text-to-Video)
3. Choose a model variant and fill in the prompt
4. Execute — the node waits for completion and returns resultUrls

Fire-and-forget: Disable Wait for Completion and use Query Task Status later with the returned taskId.

Rate Limits

Kie.ai enforces 20 requests per 10 seconds. When throttled, the API returns HTTP 429.

Automatic retry with exponential backoff:

| Retry | Delay | |-------|-------| | 1st | 2 s | | 2nd | 4 s | | 3rd | 8 s |

After 3 retries the error is propagated. For high-volume workflows, use waitForCompletion: false and manage concurrency at the n8n level.

Links

• Kie.ai
• Kie.ai API Docs
• npm package
• GitHub Repository

Changelog

See CHANGELOG.md for the full history.

| Version | Highlights | |---------|-----------| | v0.13.0 | ✨ NEW: Claude node (Opus 4.6, Sonnet 4.6, Haiku 4.5), OpenAI Chat (GPT-5.2, Codex), Midjourney (Niji 7); Suno V5.5 + Sounds + Mashup operations; Grok multi-image support; auto-publish CI | | v0.12.2 | Fix Kling 3.0 multi_shots cannot be empty 422 — different API schema from v2.x now fully handled | | v0.12.1 | Fix Kling 3.0 mode field always required — fixes 500 error | | v0.12.0 | ✨ NEW: Gemini node (2.5 Flash/Pro, 3 Pro, 3.1 Pro) with multimodal + Google Search | | v0.11.0 | Deep audit: negativePrompt, seed, numOutputs added across 9 nodes (Hailuo, GrokImagine, Kling, Seedream, Wan, Seedance, Google, Ideogram, Flux) | | v0.10.0–v0.10.4 | GrokImagine aspect_ratio fix; Kling mode/sound fix; InfineTalk registration; Suno model format fix | | v0.5.6 | Model ID audit; published as n8n-nodes-kie-pro | | v0.4.5 | 36 API snake_case naming bugs fixed across 11 nodes | | v0.4.0 | Kling 3.0 support; 429 retry with exponential backoff | | v0.1.0 | Initial release — 21 nodes |

License

MIT