Upstage Solar LLM and Embeddings nodes for n8n workflow automation. This package provides powerful AI capabilities including chat completions, embeddings generation, and document processing through n8n's visual workflow interface.
- Solar Chat Models: Use Upstage's Solar LLM (solar-mini, solar-pro, solar-pro2) for chat completions
- Embeddings Generation: Create high-quality embeddings for semantic search and vector databases
- Document Processing: Parse, OCR, classify, and extract information from documents
- LangChain Integration: Compatible nodes for advanced AI agent workflows
- Secure Authentication: Simple API key-based authentication
- Batch Processing: Efficient batch processing for embeddings and document operations
- Flexible Input: Support for single text, batch processing, binary files, and URLs
- n8n: Version 1.0.0 or later
- Node.js: Version 18.0.0 or later
-
Enable Community Nodes (if not already enabled):
export N8N_COMMUNITY_NODES_ENABLED=true n8n start -
Install via n8n UI:
- Open n8n in your browser
- Navigate to Settings → Community Nodes
- Click Install a community node
- Enter:
n8n-nodes-upstage - Click Install
npm install n8n-nodes-upstageThen enable community nodes and restart n8n:
export N8N_COMMUNITY_NODES_ENABLED=true
n8n start- Sign up at Upstage Console
- Navigate to API Keys section
- Create a new API key
- Copy your API key
- In n8n, go to Credentials → Create New
- Search for "Upstage API"
- Enter your API key
- Click Test to verify the connection
- Click Save
- Create a new workflow in n8n
- Click Add Node and search for "Upstage"
- Select any Upstage node (e.g., "Upstage Solar Chat")
- Configure the node with your credentials
- Set up your workflow and execute!
Use Upstage Solar LLM models for chat completions with conversation support.
Supported Models:
solar-mini- Fast and efficient for basic taskssolar-pro- Powerful model for complex taskssolar-pro2- Latest and most advanced Solar model with JSON support
Key Features:
- Message-based conversation format (system, user, assistant roles)
- Configurable parameters: temperature, max tokens, top-p
- Streaming response support
- JSON output formats (solar-pro2 only)
- Reasoning effort control
- Frequency and presence penalty
Example Use Cases:
- Customer support chatbots
- Content generation
- Code generation and explanation
- Data extraction and analysis
Generate high-quality embeddings using Solar embedding models for semantic search and similarity matching.
Supported Models:
embedding-query- Optimized for search queries and questionsembedding-passage- Optimized for documents and passages
Key Features:
- Single text or batch processing (up to 100 texts per request)
- Input from node parameters or previous node data
- High-dimensional vector outputs
- Token limits: Max 204,800 total tokens, 4,000 per text (optimal: under 512)
Example Use Cases:
- Semantic search
- Document similarity matching
- Clustering and classification
- Recommendation systems
Convert documents into structured HTML/Markdown format with layout preservation.
Supported Models:
document-parse- Recommended stable modeldocument-parse-nightly- Latest experimental features
Key Features:
- Sync and async document processing
- Multiple output formats (HTML, Markdown, Text)
- OCR support with auto/force modes
- Chart recognition and table merging
- Base64 encoding for figures, tables, equations, charts
- Coordinate information inclusion
Supported Formats:
- Images: JPEG, PNG, BMP, TIFF, HEIC
- Documents: PDF, DOCX, PPTX, XLSX
Example Use Cases:
- Document digitization
- Content extraction from PDFs
- Table extraction and conversion
- Document structure analysis
Extract text from document images and PDFs with high accuracy.
Supported Models:
ocr- Recommended (always points to latest stable)ocr-250904- Specific version
Key Features:
- Multiple OCR models
- Schema options (Upstage, Clova, Google)
- Page-level text extraction
- Confidence scores
- Multiple return modes (full, text, pages, words, confidence)
Supported Formats:
- Images: JPEG, PNG, BMP, TIFF, HEIC
- Documents: PDF, DOCX, PPTX, XLSX, HWP, HWPX
Example Use Cases:
- Text extraction from scanned documents
- Invoice and receipt processing
- Form data extraction
- Multi-language OCR
Extract structured information from documents using custom JSON schemas.
Supported Models:
information-extract- Recommended
Key Features:
- Custom JSON schema definition
- Form-based or raw JSON schema input
- Support for nested structures
- Schema generation mode
- Binary file or URL input
Example Use Cases:
- Invoice data extraction
- Resume parsing
- Contract analysis
- Form data extraction
Classify documents into predefined categories with confidence scores.
Supported Models:
document-classify- Document classification model
Key Features:
- Custom category definitions
- Binary file or URL input
- Form-based or JSON schema input
- Confidence scores for classifications
Example Use Cases:
- Document type classification
- Spam detection
- Content categorization
- Quality control
LangChain-compatible chat model for use with n8n AI Chain and AI Agent nodes.
Key Features:
- Automatic model selection from API
- Token usage tracking
- Streaming support
- N8n tracing integration
- Proxy support
Connections:
- Connects to:
AI Chain,AI Agentnodes - Output: Language Model
Example Use Cases:
- AI agent workflows
- Multi-step reasoning
- Tool-using agents
- Complex AI chains
LangChain-compatible embeddings model for vector databases and AI agent workflows.
Key Features:
- LangChain Embeddings interface
- Batch processing support
- Vector database integration
- Automatic text preprocessing
Connections:
- Connects to:
AI Vector Storenodes - Output: Embeddings
Example Use Cases:
- RAG (Retrieval-Augmented Generation) workflows
- Vector database population
- Semantic search in AI agents
- Document similarity in agents
- Add Upstage Solar Chat node
- Configure with your Upstage API credentials
- Set model to
solar-mini - Add a message with role "user" and your prompt
- Execute to get AI response
- Add Upstage Embed node
- Configure credentials
- Choose appropriate model (query vs passage)
- Input your text
- Get embedding vectors for similarity search, clustering, etc.
Use the Upstage Embed node with "Array of Texts" input type to process multiple texts efficiently in a single API call.
- Read Binary File node → Read PDF
- Upstage Document Parse node → Parse to Markdown
- Process the structured output
- Upstage Solar Chat for Agent → Configure model
- AI Agent node → Connect to chat model
- Upstage Embed for Agent → Connect to vector store
- Build complex AI workflows
- Chat:
https://api.upstage.ai/v1/chat/completions - Embeddings:
https://api.upstage.ai/v1/embeddings - Document Parsing:
https://api.upstage.ai/v1/document-digitization - Document OCR:
https://api.upstage.ai/v1/ocr - Information Extraction:
https://api.upstage.ai/v1/information-extraction - Document Classification:
https://api.upstage.ai/v1/document-classify
Please refer to Upstage API Documentation for current rate limits and usage guidelines.
- Embeddings:
- Max 100 strings per request
- Max 204,800 total tokens per request
- Max 4,000 tokens per text (optimal: under 512 tokens)
- Chat:
- Max tokens vary by model (check model documentation)
- Context windows: 2,048 to 32,768 tokens depending on model
- Upstage Chat API Documentation
- Upstage Embeddings API Documentation
- Upstage Document APIs Documentation
- n8n Community Nodes Documentation
- Verify community nodes are enabled:
N8N_COMMUNITY_NODES_ENABLED=true - Restart n8n completely
- Clear browser cache
- Check n8n logs for installation errors
- Verify the package is properly installed:
npm list n8n-nodes-upstage
- Invalid API Key: Verify your API key is correct and active
- Rate Limit: Check if you've exceeded API rate limits
- Network Issues: Verify network connectivity to
api.upstage.ai - Model Availability: Ensure the selected model is available in your region
Error: "At least one message is required"
- Ensure you've added at least one message in the chat node
Error: "No input text provided"
- Check that text input is not empty
- Verify the text field name matches your input data
Error: "Invalid JSON schema"
- Validate your JSON schema syntax
- Ensure the schema follows the required format
Document Processing Fails
- Verify the file format is supported
- Check file size limits
- Ensure binary property name is correct
# Clone the repository
git clone https://github.com/UpstageAI/n8n-nodes-upstage.git
cd n8n-nodes-upstage
# Install dependencies
npm install
# Build the project
npm run build
# Run in watch mode
npm run dev# Build and create npm link
npm run build
npm link
# In your n8n installation directory
cd /path/to/n8n
npm link n8n-nodes-upstage --legacy-peer-deps
# Start n8n with community nodes enabled
N8N_COMMUNITY_NODES_ENABLED=true n8n startContributions are welcome! Please feel free to submit issues and pull requests.
- Fork the repository
- Create your feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add some amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
MIT License - see LICENSE file for details.
- Upstage Console
- Upstage Documentation
- Solar LLM Documentation
- n8n Documentation
- n8n Community Nodes
- GitHub Repository
- Report Issues
- Built with n8n workflow automation platform
- Powered by Upstage Solar AI models
- Uses LangChain for AI agent integration
Made with ❤️ by Upstage