Description
n8n-nodes-chutes
!Chutes.ai
!n8n
!License
!Tests
!Node
This is an n8n community node that provides complete access to Chutes.ai’s AI services, replicating all features available in the Chutes.ai playground including text generation, image generation, and custom inference.
Chutes.ai is an advanced AI platform that provides state-of-the-art models for various AI tasks with a focus on performance and reliability.
Installation
Community Node (Recommended)
1. Go to Settings > Community Nodes in your n8n instance
2. Search for n8n-nodes-chutes
3. Click Install
Manual Installation
npm install n8n-nodes-chutes
For self-hosted n8n instances:
cd ~/.n8n
npm install n8n-nodes-chutes
Authentication (With Example)
To use this node, you’ll need an API key from Chutes.ai:
1. Sign up at chutes.ai
2. Navigate to your API Keys dashboard
3. Create a new API key (you will only be shown it once)
4. Add the key to your n8n credentials:
– Go to Credentials > New
– Select Chutes API
– Enter your API key
– Choose your environment (Production/Sandbox)
– Save
!docs/AI-AGENT-INTEGRATION.md for detailed examples and best practices.
Usage Examples
Basic Text Generation
Generate creative text using Chutes.ai’s LLM models:
{
"nodes": [
{
"parameters": {},
"type": "n8n-nodes-base.manualTrigger",
"typeVersion": 1,
"position": [
0,
0
],
"id": "c7bf314b-4232-4f02-9146-439e30aa7437",
"name": "When clicking ‘Execute workflow’"
},
{
"parameters": {
"chuteUrl": "https://chutes-moonshotai-kimi-k2-thinking.chutes.ai",
"operation": "chat",
"messages": {
"messageValues": [
{
"role": "system",
"content": "you are the HAL9000 before it went insane, be helpful like the computer in the movie, you are to pretend you are in the movie 2001 a space odyssey, and will not break character. if the user asks questions they are also pretending to be in the movie."
},
{
"content": "can you guess how long before we reach Jupiter, HAL?"
}
]
},
"additionalOptions": {
"maxTokens": 10000
}
},
"type": "CUSTOM.chutes",
"typeVersion": 1,
"position": [
220,
0
],
"id": "7beb97a8-a59c-4041-b7f2-6be1901c79eb",
"name": "Chutes",
"credentials": {
"chutesApi": {
"id": "Lv1zGZROIfHD4yIL",
"name": "Chutes account"
}
}
}
],
"connections": {
"When clicking ‘Execute workflow’": {
"main": [
[
{
"node": "Chutes",
"type": "main",
"index": 0
}
]
]
}
},
"pinData": {},
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "0766e26a38e4cf9836ab30499d93ee64e5c7b0eae3f1edb925460d34cf41d5c5"
}
}
Image Generation
Create stunning images from text descriptions:
Click to see full JSON example node
{
"nodes": [
{
"parameters": {},
"type": "n8n-nodes-base.manualTrigger",
"typeVersion": 1,
"position": [
0,
0
],
"id": "4bee9ada-56fe-4cc1-9490-ce87839077dd",
"name": "When clicking ‘Execute workflow’"
},
{
"parameters": {
"resource": "imageGeneration",
"chuteUrl": "https://chutes-qwen-image.chutes.ai",
"prompt": "a cat in a hat that sat on a fat stack of pancakes, maple syrup flows down the side of the pancakes",
"additionalOptions": {}
},
"type": "CUSTOM.chutes",
"typeVersion": 1,
"position": [
220,
0
],
"id": "13f40013-4da6-4bc6-b909-95e8e7b1594d",
"name": "Chutes",
"credentials": {
"chutesApi": {
"id": "Lv1zGZROIfHD4yIL",
"name": "Chutes account"
}
}
}
],
"connections": {
"When clicking ‘Execute workflow’": {
"main": [
[
{
"node": "Chutes",
"type": "main",
"index": 0
}
]
]
}
},
"pinData": {},
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "0766e26a38e4cf9836ab30499d93ee64e5c7b0eae3f1edb925460d34cf41d5c5"
}
}
Video Generation
Animate stunning images with Video using text descriptions :
Click to see full JSON example node
{
"nodes": [
{
"parameters": {},
"type": "n8n-nodes-base.manualTrigger",
"typeVersion": 1,
"position": [
-200,
0
],
"id": "4bee9ada-56fe-4cc1-9490-ce87839077dd",
"name": "When clicking ‘Execute workflow’"
},
{
"parameters": {
"resource": "imageGeneration",
"chuteUrl": "https://chutes-qwen-image.chutes.ai",
"prompt": "a cat in a hat that sat on a fat stack of pancakes, maple syrup flows down the side of the pancakes",
"additionalOptions": {}
},
"type": "CUSTOM.chutes",
"typeVersion": 1,
"position": [
20,
0
],
"id": "13f40013-4da6-4bc6-b909-95e8e7b1594d",
"name": "Chutes",
"credentials": {
"chutesApi": {
"id": "Lv1zGZROIfHD4yIL",
"name": "Chutes account"
}
}
},
{
"parameters": {
"resource": "videoGeneration",
"chuteUrl": "https://chutes-wan-2-2-i2v-14b-fast.chutes.ai",
"operation": "image2video",
"prompt": "the maple syrup flows down the pancakes slowly while the cat wags his tail",
"additionalOptions": {}
},
"type": "CUSTOM.chutes",
"typeVersion": 1,
"position": [
240,
0
],
"id": "6b881c17-31d5-4f2a-9553-4dbe5711d0a6",
"name": "Chutes1",
"credentials": {
"chutesApi": {
"id": "Lv1zGZROIfHD4yIL",
"name": "Chutes account"
}
}
}
],
"connections": {
"When clicking ‘Execute workflow’": {
"main": [
[
{
"node": "Chutes",
"type": "main",
"index": 0
}
]
]
},
"Chutes": {
"main": [
[
{
"node": "Chutes1",
"type": "main",
"index": 0
}
]
]
}
},
"pinData": {},
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "0766e26a38e4cf9836ab30499d93ee64e5c7b0eae3f1edb925460d34cf41d5c5"
}
}
Use your Imagination
String together as many n8n nodes as you like, for any reason, no matter how crazy. Chutes Nodes are made to work together :
!Examples Directory, you will also need FFmpeg installed to make a GIF – but that can easily be accomplished by using the docker example found in the tests folder.
Support
Development
Contributing
We welcome contributions! Please follow this workflow:
1. Fork the repository on GitHub
2. Clone your fork and create a branch from DEV:
git clone https://github.com/YOUR-USERNAME/n8n-nodes-chutes.git
cd n8n-nodes-chutes
git checkout DEV
git checkout -b feature/your-feature-name
3. Make your changes and write tests
4. Create a Pull Request to DEV (not main)
Important:
main – they will be rejectedDEV branchLocal Development Setup
Clone the repository
git clone https://github.com/chutesai/n8n-nodes-chutes.git
cd n8n-nodes-chutesCheckout DEV branch (always develop from DEV)
git checkout DEVInstall dependencies
npm installBuild the node
npm run buildLink for local development
npm link
cd ~/.n8n/custom
npm link n8n-nodes-chutesStart n8n
n8n start
Testing
This project uses Jest for testing and follows Test-Driven Development (TDD).
Run all tests
npm testRun tests with coverage
npm run test:coverageRun tests in watch mode
npm run test:watchRun specific test file
npm test -- --testPathPattern="ChutesChatModel"Run integration tests (requires CHUTESAPIKEY)
export CHUTESAPIKEY=yourapikey
npm test -- --testPathPattern="tests/integration"
Test Structure:
tests/nodes/ – Unit tests for all nodestests/integration/ – Integration tests with real API callstests/credentials/ – Credential configuration testsSee tests/README.md for detailed testing documentation.
Compatibility
Changelog
[0.1.0] – 2026-02-15 Official Release
#### Tool Calling Support (AI Agent)
role: 'tool' and toolcallid#### ChutesAIAgent & ChutesChatModel Improvements
#### Multi-Image Edit Support
#### Bug Fixes
[0.0.10] – 2026-01-14
#### LTX-2 Full Support
imagestrength, imageframe_index, distilled pipeline optionfps/framerate, frames/numframes, etc.)[0.0.9] – 2025-10-22
#### Initial Release
/v1/chat/completions for all text operationsLicense
Discord
Chutes
Author
Vonkaiser
Acknowledgments
—
Made with respect for the n8n and Chutes.ai communities
 following industry standards</li>
<li>All playground parameters:</li>
<p> – Temperature (0-2)<br />
– Max tokens<br />
– Top P (nucleus sampling)<br />
– Frequency/presence penalties<br />
– Stop sequences<br />
– Response format (text/JSON)<br />
– Streaming support (planned)<br />
– Seed for reproducibility</p>
<h3>🎨 Image Generation</h3>
<li><strong>Generate</strong>: Create images from text prompts</li>
<li><strong>Edit</strong>: Modify existing images with text prompts</li>
<p> – <strong>Multi-image composition</strong>: Combine multiple images using models like Qwen-Image-Edit-2511<br />
– Support for 1-3 images per edit (named binaries, URLs, or auto-mapped)</p>
<li>Multiple size options (256×256 to 1792×1024)</li>
<li>Quality settings (Standard/HD)</li>
<li>Style presets (Natural/Vivid)</li>
<li>Negative prompts</li>
<li>Guidance scale control</li>
<li>Seed for reproducibility</li>
<li>Batch generation support</li>
<h3>🔧 Inference (coming soon)</h3>
<li><strong>Predict</strong>: Run custom model inference</li>
<li><strong>Batch</strong>: Process multiple inputs efficiently</li>
<li><strong>Status</strong>: Check job status for long-running operations</li>
<li>Support for custom models deployed on Chutes.ai</li>
<li>Webhook callbacks for async operations</li>
<li>Priority queue management</li>
<h2>Three Ways to Use Chutes.ai with n8n</h2>
<p>This package provides three complementary nodes for different use cases:</p>
<h3>1. Chutes (Original Workflow Node)</h3>
<p><strong>Use when:</strong> You need direct control over API calls and access to all Chutes.ai features</p>
<p><strong>Features:</strong></p>
<li>📝 Text generation (completions and chat)</li>
<li>🎨 Image generation</li>
<li>🎵 Music generation</li>
<li>🎤 Speech-to-text and text-to-speech</li>
<li>🎥 Video generation</li>
<li>📊 Embeddings</li>
<li>🛡️ Content moderation</li>
<li>🔧 Custom inference</li>
<p><strong>Perfect for:</strong> Complex workflows that require full control over Chutes.ai API parameters</p>
<h3>2. Chutes Chat Model</h3>
<p><strong>Use when:</strong> You want to use Chutes.ai with an available AI Agent node</p>
<p><strong>Features:</strong></p>
<li>✅ Works with Chutes AI Agent</li>
<li>✅ <strong>Full tool calling support</strong> – passes tools to API correctly</li>
<li>✅ All Chutes LLM models (DeepSeek, Qwen, etc.)</li>
<li>✅ Dynamic chute/model selection</li>
<li>✅ Temperature and advanced parameters</li>
<li>✅ Compatible with AI Agent’s Tools, Memory, and Output Parsers</li>
<p><strong>Quick Start:</strong><br />
1. Add a <strong>Chutes Chat Model</strong> node<br />
2. Select your chute and model<br />
3. Connect to <strong>AI Agent</strong> node’s model input</p>
<h3>3. Chutes AI Agent (New!)</h3>
<p><strong>Use when:</strong> You want n8n’s full AI Agent experience, but exclusively with Chutes.ai models</p>
<p><strong>Features:</strong></p>
<li>🤖 Functions identically to n8n’s AI Agent</li>
<li>🔒 <strong>Whitelisted</strong> – Only accepts Chutes Chat Model (ensures Chutes.ai integration)</li>
<li>🛠️ <strong>Full tool calling support</strong> – Works with n8n built-in tools (Wikipedia, Calculator, SerpAPI, Code Tool, HTTP Request, Workflow)</li>
<li>📞 <strong>OpenAI-compatible function calling</strong> – Works with DeepSeek, Qwen, and other models supporting tools</li>
<li>🎯 Direct chute selection (no separate Chat Model node required)</li>
<li>💬 Multi-turn conversations with context</li>
<li>🎯 System message configuration</li>
<li>🔄 Max iterations control</li>
<li>📊 Intermediate steps tracking</li>
<p><strong>Perfect for:</strong> Full AI Agent workflows that must use Chutes.ai models only</p>
<p><strong>Quick Start:</strong><br />
1. Add <strong>Chutes AI Agent</strong> node<br />
2. Configure your chute directly (or connect a Chutes Chat Model node)<br />
3. (Optional) Add Tools, Memory, Output Parser<br />
4. Configure prompt and system message<br />
5. Execute!</p>
<p><strong>Why This Node?</strong> n8n’s official AI Agent has a hardcoded whitelist that can limit community model integration. The Chutes AI Agent provides the same functionality while ensuring seamless Chutes.ai integration.</p>
<h3>Which Node Should You Use?</h3>
<p>| Use Case | Recommended Node |<br />
|———-|—————–|<br />
| Complex workflows with multiple AI operations | <strong>Chutes</strong> (Original) |<br />
| Using Chutes official AI Agent with Chutes models | <strong>Chutes Chat Model</strong> |<br />
| AI Agent that ONLY works with Chutes models | <strong>Chutes AI Agent</strong> |<br />
| Image/video/audio generation | <strong>Chutes</strong> (Original) |<br />
| Custom model inference | <strong>Chutes</strong> (Original) |<br />
| Embeddings and content moderation | <strong>Chutes</strong> (Original) |<br />
| Tools, Memory, Output Parser with Chutes | <strong>Chutes AI Agent</strong> |</p>
<h3>Three Nodes, One Package</h3>
<p>When you install <code>n8n-nodes-chutes</code>, you get all three nodes:</p>
<p>1. <strong>Chutes</strong> – Traditional workflow node for direct API calls to all Chutes.ai features<br />
– Text generation, image generation, speech, video, embeddings, etc.<br />
– Full control over API parameters<br />
– Multi-image edit support for composition workflows</p>
<p>2. <strong>Chutes Chat Model</strong> – LangChain-compatible chat model node<br />
– Connects to Chutes official AI Agent<br />
– Provides Chutes.ai LLM models to any AI Agent<br />
– Full tool calling support (OpenAI-compatible)<br />
– <code>NodeConnectionType.AiLanguageModel</code> output</p>
<p>3. <strong>Chutes AI Agent</strong> – Complete AI Agent with Chutes-only whitelist<br />
– Clone of n8n’s AI Agent structure<br />
– <strong>Direct chute selection</strong> (no separate Chat Model required)<br />
– <strong>Full tool calling support</strong> with LangChain tools<br />
– Full Tool, Memory, and Output Parser support<br />
– Perfect for workflows that must use Chutes.ai exclusively</p>
<p>Use the right node for your use case (see the comparison table above).</p>
<p>📚 <strong>Full Guide</strong>: See <a href=){kind=link}
</h3>
<p>The node automatically handles Chutes.ai rate limits with exponential backoff and retry logic, ensuring reliable operation even under heavy load.</p>
<h3>🛡️ Error Handling</h3>
<p>Comprehensive error messages help you debug issues quickly with detailed Chutes.ai API error information and suggestions.</p>
<h3>📦 Batch Processing</h3>
<p>Process multiple items efficiently with automatic batching for optimal API usage and cost management.</p>
<h2>Playground Feature Parity</h2>
<p>This node aims for complete feature parity with the Chutes.ai playground:</p>
<li>✅ All models available in the playground</li>
<li>✅ All parameters and options</li>
<li>✅ Same response formats</li>
<li>✅ Identical error handling</li>
<li>✅ Real-time streaming</li>
<li>✅ Dynamic model loading</li>
<h2>Common Workflows</h2>
<h3>Content Generation Pipeline</h3>
<p>1. <strong>Chutes Text Generation</strong> → Generate article outline<br />
2. <strong>Chutes Text Generation</strong> → Expand each section<br />
3. <strong>Chutes Image Generation</strong> → Create featured image<br />
4. <strong>Save to Database/CMS</strong></p>
<h3>AI-Powered Customer Support</h3>
<p>1. <strong>Webhook</strong> → Receive customer inquiry<br />
2. <strong>Chutes Chat</strong> → Generate response with context<br />
3. <strong>Chutes Text Generation</strong> → Summarize conversation<br />
4. <strong>Send Response</strong></p>
<h3>Image Processing Pipeline</h3>
<p>1. <strong>Trigger</strong> → New product added<br />
2. <strong>Chutes Image Generation</strong> → Create product images<br />
3. <strong>Chutes Text Generation</strong> → Generate descriptions<br />
4. <strong>Upload to Storage</strong></p>
<h2>Technical Details</h2>
<h3>Chat Completions API</h3>
<p>This node uses modern best practices by implementing the Chat Completions API (<code>/v1/chat/completions</code>) for <strong>all text generation operations</strong>:</p>
<li><strong>Complete Operation</strong>: Internally wraps your prompt as a user message for the Chat Completions API</li>
<li><strong>Chat Operation</strong>: Uses the full messages array with system, user, and assistant roles</li>
<li><strong>Why This Matters</strong>:</li>
<p> – ✅ Follows industry standards (OpenAI deprecated <code>/v1/completions</code> in favor of chat)<br />
– ✅ Better quality responses (how modern LLMs are trained)<br />
– ✅ Consistent behavior across all text operations<br />
– ✅ No truncation issues (the <code>/v1/completions</code> endpoint has known truncation problems on Chutes.ai)</p>
<p><strong>User Experience</strong>: The UI remains simple – you still enter a prompt for the “Complete” operation. The conversion to chat format happens automatically behind the scenes!</p>
<h3>Dynamic Chute Loading</h3>
<p>The node dynamically loads available chutes from the Chutes.ai Management API:</p>
<li>Fetches all public and your deployed chutes</li>
<li>Filters by resource type (LLM, Image, Video, Audio, TTS, STT, Music, Embeddings, Moderation)</li>
<li>Each chute has its own subdomain: <code>https://{slug}.chutes.ai</code></li>
<li>Automatically routes requests to the correct endpoint</li>
<h3>Error Handling & Retries (better features coming soon)</h3>
<li>Automatic retry logic with exponential backoff (coming soon)</li>
<li>Handles rate limiting (429 errors)</li>
<li>Clear error messages for authentication and API issues</li>
<li>Graceful degradation for optional features</li>
<h2>Resources</h2>
<li><a href=){kind=link}