n8n-nodes-kaia

n8n community node package for interacting with the Kaia API.

Nodes

Save Signal

Capture a signal for knowledge graph processing. This node automatically populates workflow context fields and sends signals to the Kaia API.

Features:

• Auto-populates workflow context (workflowId, workflowName, executionId, nodeId, nodeName)
• Supports all signal types and optional fields
• Auto-generates SHA-256 content hash for deduplication
• Supports LLM context, knowledge graph, tags, and priority levels

Security Gate

Check entity context from the knowledge graph and block workflow execution if conditions are not met.

Features:

• Evaluates entity facts and relations
• Supports multiple operators (equals, not equals, exists, not exists)
• Blocks workflow execution when conditions are not met
• Returns full context data when proceeding

Model Gate

Intercept LLM input and output to record all interactions for AI Act compliance. This node can be placed after any LLM node to capture and log AI interactions.

Features:

• Captures LLM prompts, responses, and metadata
• Extracts tool calls and decisions from AI responses
• Records token usage and latency metrics
• Sends separate input and output signals to Kaia API
• Non-blocking: workflow continues even if signal sending fails
• Includes signal status in output for debugging (_kaiaSignalStatus)

Usage:

1. Place the Model Gate node after an LLM node (e.g., OpenAI, Anthropic, etc.)
2. Configure Kaia API credentials
3. The node will automatically capture input/output and send compliance signals

Compatibility:

• Tested with n8n v1.x+
• Works with all LangChain-compatible LLM nodes

Credentials

Kaia API

Bearer token authentication with base URL configuration.

Required Fields:

• Base URL: The base URL of the Kaia API
• Bearer Token: Authentication token

Installation

npm install n8n-nodes-kaia

Development

# Install dependencies
npm install

# Build the package
npm run build

# Run linter
npm run lint

# Format code
npm run format

License

MIT