n8n has two built-in Vertex AI nodes for its AI features: the Google Vertex Chat Model for AI agents and chains, and Embeddings Google Vertex. Vertex AI community nodes add what those leave out: Anthropic’s Claude models hosted on Vertex AI, Gemini as a regular action node, image and video generation, context caching and batch jobs. This guide compares them, using n8n’s documentation, the description and README each author publishes on npm and npm’s download counts.
The built-in Vertex AI nodes
According to n8n’s documentation, the Google Vertex Chat Model takes a model name such as gemini-2.5-flash, a region (the credential’s, or Global, EU or US multi-region) and options for maximum tokens, temperature, a thinking budget, Top K, Top P and Gemini safety settings. Embeddings Google Vertex generates embeddings with the model you name. Both sign in with a Google Vertex AI credential or a Google service account credential: you need a Google Cloud project with billing and the Vertex AI API enabled, and a service account with the roles/aiplatform.user role whose JSON key gives the credential’s email and private key.
Vertex AI community nodes
Our Vertex AI hub lists 9 packages that pass our quality rule. The most downloaded, with the description each author gives on npm:
| Package | npm downloads, 20 to 26 October 2026 | Latest version | Author’s description |
|---|---|---|---|
| n8n-nodes-mcp-vertex-safe | 1,149 | 0.3.1 (11 August 2026) | An MCP Client Tool that cleans tool schemas for Gemini, and a chat model that calls Claude through Vertex Model Garden |
| n8n-nodes-vertex-ai | 1,039 | 0.2.0 (18 September 2026) | Google Vertex AI (Gemini) |
| @brunodosanjos/n8n-nodes-vertex-ai-pdf | 1,024 | 1.1.0 (31 January 2026) | Google Vertex AI (Gemini) |
| @schamardak/n8n-nodes-vertex-anthropic | 954 | 0.1.0 (19 May 2026) | Anthropic Claude chat model through Google Cloud Vertex AI (Model Garden) |
| n8n-nodes-google-vertex-ai | 86 | 0.1.3 (9 June 2026) | Gemini text, multimodal analysis, Nano Banana image generation and editing, and Veo video |
| n8n-nodes-prewave-vertex-ai | 74 | 0.1.0 (1 June 2026) | Gemini and Anthropic Claude through a stored Google service account |
| @boazlai/n8n-nodes-vertex-ai | 74 | 0.1.0 (30 April 2026) | Gemini context caching, batch prediction, GCS and a chat model with cached content |
Versions and dates come from the npm registry. Downloads are npm’s raw count for the week of 20 to 26 October 2026. For the first four packages, most of that week’s downloads came on a single day, 24 October, so their weekly figures are irregular spikes rather than a steady install rate.
Claude on Vertex AI
Two READMEs explain why separate nodes exist: n8n’s built-in Google Vertex Chat Model sends Gemini-format requests, so it can’t call Claude models from Vertex Model Garden, and the built-in Anthropic chat model calls Anthropic’s own API rather than Vertex.
- @schamardak/n8n-nodes-vertex-anthropic: a chat model sub-node for the AI Agent with tool calling, for Claude models hosted on Vertex AI in the region you choose. Its README says it was built to keep the model calls in Vertex AI’s EU regions for data residency.
- n8n-nodes-mcp-vertex-safe: has a similar Anthropic Chat Model (Vertex), plus an MCP Client Tool (Vertex-Safe). Its README describes a
400 Bad Requestwhen an AI Agent uses Vertex Gemini with MCP Client Tool nodes attached, because Vertex accepts only a restricted subset of JSON Schema; this node cleans the MCP tool schemas before sending them. It links the two n8n GitHub issues it tracks. - n8n-nodes-prewave-vertex-ai: one service-account credential for both Gemini and Claude on Vertex AI, with an action node (chat and generate, embeddings and Imagen images) and a chat model sub-node with a provider toggle.
Gemini on Vertex AI as an action node
- n8n-nodes-vertex-ai: text generation, multi-turn chat, images as input and structured JSON or enum output with Gemini models. @brunodosanjos/n8n-nodes-vertex-ai-pdf has the same description and author name on npm, and no README.
- n8n-nodes-google-vertex-ai: Gemini text and multimodal analysis, image generation and editing with the Gemini image models, and Veo video generation, with a service-account credential. Its README states that it is unofficial and not affiliated with Google.
- @boazlai/n8n-nodes-vertex-ai: Gemini context caches (create, get, list, update TTL and delete), batch prediction jobs, Cloud Storage uploads and downloads, and a chat model sub-node that can use a cached context.
Check each node page here for the latest release and, where it has one, its GitHub repository for open issues before you build on a package.
How to choose
- Gemini in an AI agent, or embeddings for a vector store: the built-in Google Vertex Chat Model and Embeddings Google Vertex nodes, with nothing to install.
- Claude through your Google Cloud account: @schamardak/n8n-nodes-vertex-anthropic or n8n-nodes-mcp-vertex-safe, or n8n-nodes-prewave-vertex-ai for Gemini and Claude under one credential.
- MCP tools with a Vertex Gemini agent that fail with 400 Bad Request: the MCP Client Tool (Vertex-Safe) in n8n-nodes-mcp-vertex-safe.
- Images or video on Vertex AI: n8n-nodes-google-vertex-ai.
- Large prompts reused often, or batch jobs: @boazlai/n8n-nodes-vertex-ai.
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: AWS and n8n.
FAQ
Does n8n have a Vertex AI node?
Yes, two built-in nodes for its AI features: the Google Vertex Chat Model and Embeddings Google Vertex. Community nodes add Claude on Vertex AI, action nodes, image and video generation, and caching and batch jobs.
How do I connect n8n to Vertex AI?
Enable the Vertex AI API in a Google Cloud project with billing, create a service account with the roles/aiplatform.user role and a JSON key, and enter its email, private key, project ID and region in a Google Vertex AI credential in n8n.
Can I use Claude on Vertex AI in n8n?
Not with the built-in Google Vertex Chat Model, which sends Gemini-format requests according to these READMEs. @schamardak/n8n-nodes-vertex-anthropic, n8n-nodes-mcp-vertex-safe and n8n-nodes-prewave-vertex-ai call Claude on Vertex AI.
What is the most downloaded Vertex AI community node?
n8n-nodes-mcp-vertex-safe, with 1,149 npm downloads in the week of 20 to 26 October 2026, most of them on 24 October.
