n8n-nodes-kieai

n8n community node package integrating multiple Kie.ai image and video models.

Overview
- Models: 4o Image, Flux.1 Kontext, Qwen Image Edit, Google Imagen 4, Ideogram v3, Ideogram Character, Midjourney, WAN 2.2, Runway (AI Video + Aleph), Kling v2.1, Seedance, Veo 3 (incl. 1080p), Suno (Music), Luma Modify (Video), plus Upload and Account (credits).
- Features: async polling and callbacks, result URL extraction, optional binary download, simple upload helper.

Install (n8n UI)
- Settings > Community Nodes > Install
- Search for n8n-nodes-kieai and install
- Create credentials "Kie.ai API" and paste your API key

Install (Build & Link)
``powershell
cd C:\Users\Igor\Desktop\n8n-nodes-kieai
if (Test-Path .\dist) { Remove-Item -Recurse -Force .\dist }
npm run build
npm link

cd $env:USERPROFILE\.n8n\custom
npm link n8n-nodes-kieai
npx n8n start
`

Credentials
- In n8n, create a credential: "Kie.ai API" and paste your API key.
- The node attaches Authorization: Bearer to all Kie.ai requests.

Usage
- Add the "Kie.ai" node from the sidebar.
- Choose a resource (model) and an operation (e.g., generate, details, downloadUrl).
- For long-running jobs, either return taskId immediately or enable polling and optionally download the first result as binary.

Notes
- Endpoint conventions vary: most models use /record-info; some "playground" APIs use /recordInfo.
- JSON workflow import may not always map node types correctly; prefer adding the node from the sidebar when testing.

Troubleshooting Import (Unknown node "?")
- Ensure this package is linked in %USERPROFILE%\.n8n\custom and n8n has been restarted.
- The expected node type in workflow JSON is n8n-nodes-kieai.kieAi.
- If an imported workflow uses a different variant (e.g., kieAi, n8n-nodes-kieai.KieAi), normalize it:
- Run: npm run fix:workflow -- path\to\workflow.json
- Re-import the adjusted JSON into n8n.

Development
- Source in nodes/ and credentials/, outputs in dist/.
- The build step copies .png/.svg logos from nodes/ to dist/nodes/.

Quickstarts by Model

4o Image (GPT-Image-1)
- Resource: 4o Image (GPT-Image-1)
- Operation: Generate
- Set Prompt; optional: Reference Image URLs (one per line or comma), Aspect Ratio, Enhance, Variants
- For immediate results: set Mode = poll, adjust Poll Interval/Max, and enable Download as Binary to attach the first image
- Extras: use Get Details to re-fetch status by taskId; use Get Direct Download URL to convert a temporary URL into a signed URL

Flux.1 Kontext
- Resource: Flux.1 Kontext
- Operation: Generate
- Set Prompt; optional: Input Image URL, Aspect Ratio, Model, Output Format, Translation, Prompt Upsampling, Safety, Callback
- Use Mode = poll to wait; optionally download first result as binary

Qwen Image Edit
- Resource: Qwen Image Edit
- Operation: Generate (edit)
- Provide Prompt and one or more Image URLs
- Use Mode = poll to wait; optionally download first result
- Use Details with a taskId to re-check status

Google Imagen 4
- Resource: Google Imagen 4
- Operation: t2i for text-to-image, i2i for image-to-image
- Set Prompt (and input image for i2i); optional: Negative Prompt, Aspect Ratio, Output Format, Steps, Guidance, Seed, Safety, Callback
- Mode = poll for immediate results; optionally download first result
- Use Details to re-fetch by taskId

Ideogram v3
- Resource: Ideogram v3
- Operation: Generate (and Edit/Remix if present)
- Set Prompt; optional: variant settings per UI
- Mode = poll to wait; optionally download first result; use Details for status

Ideogram Character
- Resource: Ideogram Character
- Operation: Generate (and Edit/Remix if present)
- Similar to Ideogram v3; Mode = poll for immediate results; use Details for status

Midjourney
- Resource: Midjourney
- Operations: Generate, Upscale, Vary, Extend Video, Details
- Start with Generate + Prompt; poll for result; then apply Upscale/Vary using returned task details

WAN 2.2 (Video)
- Resource: WAN 2.2 (Video)
- Operations: t2v (text->video), i2v (image->video), Details
- For t2v: Prompt, Resolution, Aspect Ratio; optional: Prompt Expansion, Seed, Acceleration, Callback
- For i2v: Image URL + Prompt, Resolution, Aspect Ratio; same options
- Mode = poll; optionally enable Download as Binary to fetch the first video URL

Runway (AI Video)
- Resource: Runway (AI Video)
- Operations: t2v, i2v, extend, details
- Start with t2v/i2v + Prompt (and Image URL for i2v); Mode = poll; optionally download first result; use extend to lengthen a video by taskId

Runway Aleph (Video)
- Resource: Runway Aleph (Video)
- Similar to Runway (AI Video), targeting Aleph models; poll for results and use Details by taskId as needed

Kling v2.1 (Video)
- Resource: Kling v2.1 (Video)
- Operation: Generate (per UI)
- Provide Prompt and any model-specific options shown; Mode = poll; optionally download first result; use Details for status

Seedance v1 (Video)
- Resource: Seedance v1 (Video)
- Operation: Generate
- Provide Prompt and options; Mode = poll; optionally download first result; use Details for status

Veo 3 (Quality / Fast)
- Resource: Veo 3 (Quality / Fast)
- Operations: Generate, Details, Get 1080p
- Generate: set Model (fast/quality), Prompt, Aspect Ratio; optional: image URLs, watermark, seeds, fallback, callback; Mode = poll
- After completion: use Get 1080p with taskId and index to fetch the full HD URL if provided

Nano Banana
- Resource: Nano Banana (Generate/Edit)
- Operations: Generate (text->image), Edit (image->image), Details
- Generate: Prompt; optional Output Format and Image Size; Mode = poll
- Edit: Edit Prompt + one or more Image URLs; Mode = poll
- Details: re-check using taskId; status path uses /playground/recordInfo

Suno (Music)
- Resource: Suno (Music)
- Operations: Generate Music, Generate Lyrics, Details
- Generate Music: set Prompt; optional Lyrics/Style/Duration/Callback; Mode = poll to wait; optionally download first audio as binary
- Details: re-check using taskId

Luma Modify (Video)
- Resource: Luma Modify (Video)
- Operations: Modify, Details
- Modify: set Prompt; optional Source Video URL/Strength/Aspect/Duration/Callback; Mode = poll to wait; optionally download first video URL as binary
- Details: re-check using taskId

Upload
- Resource: Upload
- Operations: From Binary (input) or From URL
- From Binary: set Binary Property (default data), optional File Name/Content Type
- From URL: provide File URL, optional File Name/Content Type overrides
- Returns a downloadUrl` for use with other models

Account
- Resource: Account
- Operation: Get Credits — returns your Kie.ai credit balance