n8n has two built-in nodes for Anthropic’s Claude models: the Anthropic app node, a normal workflow step for messages, documents, images, files and prompts, and the Anthropic Chat Model sub-node, which gives an AI Agent its model. Both use one Anthropic credential. The Claude community nodes go in other directions: Claude Code and the Claude CLI, Claude on Google Vertex AI, and more of the Anthropic API. This guide covers the built-in nodes (from n8n’s documentation) and the most used Claude community packages (from our own npm data).
Which Claude node do you need?
| You want to | Use |
|---|---|
| Message Claude, analyze documents or images, manage files, or generate and improve prompts, as a workflow step | Anthropic node (built in) |
| Give an AI Agent a Claude model, with prompt caching | Anthropic Chat Model (built in, sub-node) |
| Run Claude Code or the Claude CLI, use Claude on Vertex AI, or use API features the built-in nodes don’t expose | A community node (see below) |
Set up the Anthropic credential
You need an Anthropic Console account with access to Claude. Then:
- In the Anthropic Console, open Settings > API Keys.
- Select + Create Key and give it a name, such as
n8n-integration. - Select Copy Key and enter it as the API Key in n8n.
- Optional: turn on Add Custom Header and enter a header name and value to send with every request.
On n8n Cloud you can skip the credential: select Use Gateway credits in the credential field and the node runs without an Anthropic account.
The Anthropic node
- Document: Analyze Document (answer questions about documents)
- File: Upload File, Get File Metadata, List Files, Delete File
- Image: Analyze Image (answer questions about images)
- Prompt: Generate Prompt, Improve Prompt, Templatize Prompt
- Text: Message a Model (create a completion)
The Anthropic Chat Model sub-node
Use it to give Claude models to a conversational agent. Pick the Model, then set the options:
- Maximum Number of Tokens: the completion length.
- Sampling Temperature: higher values give more varied output, with a higher risk of hallucinations.
- Top K: how many token choices the model considers for the next token.
- Top P: lower values ignore less probable options.
- Prompt Caching: caches the system prompt, tool definitions and conversation history between requests. Choose Disabled (default), 5 Minutes or 1 Hour, which sets how long the cache stays valid.
Claude community nodes
Our Claude hub lists 33 packages that pass our quality rule, with 5,168 npm downloads a week between them. The most downloaded, with the description each author gives on npm (weekly npm downloads as of 27 October 2026):
| Package | Weekly downloads | Author’s description |
|---|---|---|
| n8n-nodes-claude-cli-bridge | 2,623 | The locally installed Claude CLI, signed in with a subscription; passes images and PDFs to Claude; no Anthropic API key |
| @schamardak/n8n-nodes-vertex-anthropic | 1,659 | Claude chat model through Google Cloud Vertex AI (Model Garden) |
| @johnlindquist/n8n-nodes-claudecode | 105 | Claude Code SDK integration with MCP support |
| n8n-nodes-claude-code-cli | 67 | Claude Code CLI for AI-assisted coding tasks |
| n8n-nodes-claude-pro | 65 | Claude Pro via setup-token authentication, no CLI on the server |
| n8n-nodes-claude-sdk-agent | 44 | Claude Agent SDK with model selector, streaming, MCP and skills |
| @chrishdx/n8n-nodes-claude-cli | 42 | Claude CLI as a LangChain chat model |
| n8n-nodes-claude-refine | 40 | Anthropic API messages with prompt caching, extended thinking, structured outputs, citations, batches and more, plus a chat model |
| n8n-nodes-revenium-anthropic | 39 | Anthropic chat completion with Revenium usage tracking |
From these descriptions, the packages fall into three groups:
- Claude Code, the Claude CLI and the Agent SDK (claude-cli-bridge, claudecode, claude-code-cli, claude-pro, claude-sdk-agent, chrishdx claude-cli). Several of them sign in with a Claude subscription instead of an API key, and some need the CLI installed on the n8n server. Read the README for what each needs, and check that the way it signs in fits Anthropic’s terms for your account.
- Claude on another cloud: vertex-anthropic runs Claude models hosted on Google Cloud Vertex AI as a chat model.
- More of the Anthropic API: claude-refine lists extended thinking, structured outputs, citations, message batches and token counting; revenium-anthropic adds usage tracking.
Check each package’s node page here for its latest release and, where it has one, its GitHub repository for open issues.
How to choose
- Claude with an Anthropic API key, or with Gateway credits on n8n Cloud: the built-in nodes.
- Claude Code or agent-style coding tasks from a workflow: one of the Claude Code or CLI packages, on self-hosted n8n.
- Your company uses Claude through Google Cloud: the Vertex AI package.
- A specific API feature the built-in nodes don’t show: compare the README of claude-refine and similar packages.
To install a package on a self-hosted instance, an owner or admin enters its npm name in Settings > Community Nodes; full steps are in How to Install n8n Community Nodes. On n8n Cloud only verified community nodes can be installed, see Verified vs Unverified n8n Community Nodes. Related: OpenAI in n8n, DeepSeek in n8n and MCP in n8n.
FAQ
Does n8n support Claude without community nodes?
Yes. The Anthropic node and the Anthropic Chat Model are built in and use an Anthropic API key, or Gateway credits on n8n Cloud.
Does the Anthropic Chat Model support prompt caching?
Yes. Set Prompt Caching to 5 Minutes or 1 Hour to cache the system prompt, tool definitions and conversation history between requests.
Can I run Claude Code from n8n?
Not with the built-in nodes. Several community packages on our Claude hub describe Claude Code, Claude CLI or Agent SDK integration, mostly for self-hosted n8n.
What is the most used Claude community node?
n8n-nodes-claude-cli-bridge, with 2,623 weekly npm downloads on 27 October 2026.
