n8n-nodes-claude

This is an n8n community node package that provides two powerful nodes for integrating Anthropic's Claude AI into your n8n workflows:

1. Claude Node: Execute Claude CLI commands with full agentic capabilities
2. PRD Agent Node: Product Requirements Document management with specialized AI agents

n8n is a fair-code licensed workflow automation platform.

Anthropic Claude is a family of large language models developed by Anthropic.

Claude CLI is a command-line interface that provides agentic AI capabilities with file system access and tool usage.

Prerequisites

Claude CLI must be installed on the system running n8n:

``bash
npm install -g @anthropic-ai/claude-code
`

For detailed installation instructions, see the Claude CLI Installation Guide.

Installation

Follow the installation guide in the n8n community nodes documentation.

$3

1. Navigate to your n8n custom nodes directory:
`bash
cd ~/.n8n/nodes
`

2. Install dependencies:
`bash
cd n8n-nodes-claude
npm install
`

3. Build the node:
`bash
npm run build
`

4. Restart n8n to load the custom node

Nodes Overview

$3

This node executes Claude CLI commands with full agentic capabilities including file system access, tool usage, and multi-turn reasoning.

$3

Specialized node for Product Requirements Document management with five operations:
- Draft PRD: Generate initial PRD from a goal
- Review Block: Edit, confirm, or rollback PRD blocks
- Compile Spec: Generate technical specifications
- QA Guide: Refine PRD based on Q&A answers
- Generate Diagrams: Create ER and state machine diagrams

See PRD Node Documentation for detailed information and examples.

Usage

$3

This node executes Claude CLI commands with full agentic capabilities including file system access, tool usage, and multi-turn reasoning.

Parameters:
- Prompt (required): The command/query to send to Claude CLI
- CLI Options:
- Output Format: Choose JSON, Text, or Stream JSON (default: json)
- Max Turns: Maximum number of agentic iterations (default: 5)
- Verbose: Enable detailed turn-by-turn logging
- Model Alias: Choose sonnet, opus, or haiku (default: sonnet)
- Additional Directories: Comma-separated paths to make accessible
- Permission Mode: How to handle permissions (ask/allow/deny, default: allow)
- Input Text: Optional text to pipe as input to the CLI

$3

`
Prompt: "What is 2+2?"
CLI Options:
Output Format: json
Model Alias: sonnet
`

$3

`
Prompt: "Review this code and suggest improvements"
CLI Options:
Output Format: json
Max Turns: 10
Additional Directories: /path/to/project
Permission Mode: allow
Input Text: "{{ $json.code }}"
`

$3

`
Prompt: "Analyze this log file for errors"
CLI Options:
Output Format: json
Verbose: true
Input Text: "{{ $json.logContent }}"
`

$3

`
Prompt: "Refactor all components to use TypeScript strict mode"
CLI Options:
Output Format: json
Max Turns: 20
Additional Directories: /path/to/src/components
Permission Mode: allow
`

Output Format

- JSON Output: Returns structured JSON with the CLI response
- Text Output: Returns plain text in { text: "...", raw: "..." } format

Example JSON output:
`json
{
"text": "Response from Claude...",
"raw": "Full CLI output..."
}
`

Example Workflows

$3


`
GitHub Trigger → Claude CLI → Comment on PR
`
Automatically review code changes with full project context and file system access.

$3

`
Schedule Trigger → Claude CLI → Git Commit → Send Notification
`
Refactor code across multiple files with agentic capabilities.

$3

`
Read Log Files → Claude CLI → Parse Errors → Send Slack Alert
`
Analyze logs and extract actionable insights with detailed reasoning.

$3

`
Cron Trigger → Claude CLI → Save to Repository → Create PR
`
Generate or update documentation based on codebase changes.

$3

`
Webhook → Claude CLI → Write Tests → Run Tests → Report Results
`
Automatically generate comprehensive tests for new code.

$3

`
Schedule → Claude CLI → Generate Report → Send Email
`
Perform security audits with file access and multi-step analysis.

Best Practices

1. Permission Mode: Use allow for automated workflows to avoid interactive prompts
2. Max Turns: Increase for complex multi-step tasks (e.g., 10-20 for code refactoring)
3. Additional Directories: Grant access to specific project directories for file operations
4. Verbose Mode: Enable for debugging to see detailed execution logs
5. JSON Output: Use JSON format for easier parsing in n8n workflows
6. Security: Be cautious with permission mode allow` - only use in trusted environments
7. Error Handling: Enable "Continue On Fail" in node settings for robust workflows
8. Input Text: Use to pipe content from previous nodes for analysis or processing

Resources

- n8n documentation
- Claude CLI documentation
- Claude CLI installation guide
- n8n community forum

Compatibility

- Requires n8n version 0.210.0 or later
- Requires Claude CLI installed on the host system
- Compatible with all n8n deployment types (self-hosted, cloud, desktop)
- Works with Claude CLI version 1.0.0 and above

License

MIT

Support

For issues, questions, or feature requests, please create an issue in the repository.

---

Made with ❤️ for the n8n community