claude-pro

n8n node for Claude Pro via setup-token authentication. No CLI required on the server.

Package Information

Released: 7 mo ago
Downloads: 23 weekly / 232 monthly
Latest Version: 2.0.20
Author: Claude Pro Node Contributors

Available Nodes

Claude Pro LM icon
Claude Pro LM
Language Model node for Claude via setup-token authentication.

Documentation

n8n-nodes-claude-pro

An n8n community node for calling Claude (Anthropic) using a setup-token for authentication. No Claude CLI installation required on the n8n server.

Why this node?

Existing n8n community nodes for Claude require the Claude Code CLI installed on the server and use browser-based OAuth. This doesn't work for hosted/remote n8n instances where you can't run claude login interactively.

This node uses a different approach: generate a setup-token on your local machine, paste it into n8n credentials, and the node calls the Claude API directly with Bearer auth.

Prerequisites

  1. Install the Claude CLI on your local machine
  2. Log in: claude login
  3. Generate a setup token: claude setup-token
  4. Copy the token output

Installation

Community Nodes (recommended)

  1. Go to Settings > Community Nodes in your n8n instance
  2. Search for n8n-nodes-claude-pro
  3. Install

Manual Installation

cd ~/.n8n/custom
npm install n8n-nodes-claude-pro
# or link a local clone:
ln -s /path/to/claude-pro-node ~/.n8n/custom/node_modules/n8n-nodes-claude-pro

Restart n8n after installation.

Configuration

Credentials

  1. In n8n, go to Credentials > New Credential
  2. Search for "Claude Pro Setup Token"
  3. Paste your setup token from claude setup-token
  4. Click Test to verify the token works

Node Parameters

Parameter Default Description
Model Claude Sonnet 4 Choose between Opus 4, Sonnet 4, Haiku 3.5
Prompt — The message to send (required). Supports expressions.
System Prompt empty Optional system-level instructions
Max Tokens 4096 Maximum response length (1–128,000)
Temperature 1.0 Randomness control (0.0–2.0)
Streaming false Use SSE streaming (collected before output)
Extended Thinking false Enable chain-of-thought reasoning
Thinking Budget 10000 Token budget for thinking (shown when thinking enabled)

Output

Each item returns:

{
  "text": "Claude's response",
  "thinking": "Chain-of-thought (if extended thinking enabled)",
  "model": "claude-sonnet-4-20250514",
  "stopReason": "end_turn",
  "usage": {
    "inputTokens": 25,
    "outputTokens": 150
  }
}

Error Handling

The node provides clear error messages:

  • 401: Setup token is invalid or expired. Run claude setup-token again.
  • 403: Token does not have permission for the selected model.
  • 429: Rate limited. Retry after a delay.
  • 400: Forwards the Anthropic error message.
  • 500/529: Anthropic API error. Retry later.

Enable Continue On Fail in the node settings to receive errors as JSON output instead of stopping the workflow.

Development

git clone <repo-url>
cd claude-pro-node
npm install
npm run build

To test locally with n8n:

ln -s $(pwd) ~/.n8n/custom/node_modules/n8n-nodes-claude-pro
# restart n8n

License

MIT

Discussion