--- title: "Google Vertex AI MCP server for AI agents - actionlake" description: "Give Claude 1,128 Google Vertex AI actions through one MCP server: endpoints, locations. Your agent never holds the credential. 100,000 calls a month free." canonical: https://actionlake.com/connectors/google-vertex-ai updated: 2026-10-23 --- Connectors · AI # Google Vertex AI MCP server for AI agents. actionlake is an MCP gateway: one remote MCP server that connects Claude and other MCP clients to Google Vertex AI and the rest of your systems. It gives your agent 1,128 Google Vertex AI actions (565 reads, 242 writes and 321 that wait for a confirmation) and holds the Google Vertex AI credential itself, so the agent never does. It paces Google Vertex AI's limits and records every call. Connecting Google Vertex AI needs your own Google Vertex AI OAuth app and one admin sign-in. [Connect Google Vertex AI free](https://actionlake.com/console?start=sandbox) [See pricing](https://actionlake.com/pricing) 100,000 free calls a month. No card, no email, no password: press Start a free sandbox on the next screen and keep the key it shows you. `claude mcp add --transport http actionlake https://connect.actionlake.com/mcp` the first call ``` do { action: "gcp_vertex_ai.express_batch_prediction_jobs.list", intent: "check which Google Vertex AI account is connected" } # Express batch prediction jobs: list · read · 1 upstream call ``` - 1,128 actions in one signed file - 565 read look, never change - 242 write create and update - 321 wait for a yes destructive, confirmed first Set up ## Connect Google Vertex AI in three steps. 1. 01 ### Add actionlake to Claude. Run the command once in Claude Code; its consent page can create a free sandbox: 100,000 calls a month, no card. In the Claude app, add a custom connector with the URL below and paste your key on the consent page. Creating a new sandbox from the Claude app is not yet tested by us. `https://connect.actionlake.com/mcp` 2. 02 ### Connect Google Vertex AI. OAuth 2.0, authorization code with PKCE. You register your own OAuth app with Google Vertex AI, an admin signs in once, and you paste the app's client id and secret and the tokens from that sign-in; actionlake renews the access token from then on. actionlake stores what you paste write-only, sealed under your tenant's key. Before the first sign-in, Google Vertex AI needs an OAuth app of your own, registered in its developer portal. actionlake has no Google Vertex AI app to lend you. - `access_token` Access token from one sign-in with your own OAuth app - `refresh_token` Refresh token from the same sign-in; the access token is renewed from it - `client_id` Google OAuth client id (Google Cloud console > APIs & Services > Credentials; enable the Vertex AI API on the project) - `client_secret` Google OAuth client secret - `host` Where your Google Vertex AI lives: your host ``` account { op: "connections.requirements", service: "gcp_vertex_ai" } # free: exactly what to paste, and what is found for you ``` 3. 03 ### Make the first call. Check before a write. Ask in plain words; Claude picks the action. Before anything changes, `check` validates the call and prices it, for free. ``` do { action: "gcp_vertex_ai.express_batch_prediction_jobs.list" } # express batch prediction jobs: list · read · 1 upstream call check { action: "gcp_vertex_ai.endpoints.predict", input: { … } } # free: inputs, permissions, cost and confirmations, nothing sent ``` Google Vertex AI at a glance **Sign-in** OAuth 2.0, authorization code with PKCE. You register your own OAuth app with Google Vertex AI, an admin signs in once, and you paste the app's client id and secret and the tokens from that sign-in; actionlake renews the access token from then on. **You provide** `access_token` `refresh_token` `client_id` `client_secret` **Setup fields** `host` **Limits we pace to** 600 calls a minute, per connection **API covered** Google Vertex AI (Gemini generation, model endpoints, model registry, datasets, training, pipelines, RAG, Agent Engine) API aiplatform v1 (Discovery revision 20261012)`aiplatform.googleapis.com` **Built from** Google Vertex AI's documentation, read 22 October 2026[aiplatform.googleapis.com](https://aiplatform.googleapis.com/$discovery/rest?version=v1) [cloud.google.com](https://cloud.google.com/vertex-ai/docs/reference/rest) **Service file** Version 1.0.0, signed (Ed25519) **Cached reads** 217 of 565 reads, for up to 5 minutes; a write clears what it affects **Shared entities** 2, so ask can line Google Vertex AI up with your other systems **Data** Processed and stored in Sydney, Australia. No US region yet Capabilities ## What an AI agent can do in Google Vertex AI. 1,128 actions in 218 areas. Every one carries its effect, so a key can allow the reads and hold the rest. - **565** read look, never change - **242** write create and update - **321** destructive wait for a confirmation 1. ### Endpoints 21 5 read · 14 write · 2 destructive Endpoints: generate content · Endpoints: predict · Endpoints: raw predict · Endpoints: compute tokens 2. ### Locations 18 5 read · 13 write · 0 destructive Locations: list · Locations: get · Locations: ask contexts · Locations: async retrieve contexts 3. ### Models 15 6 read · 7 write · 2 destructive Models: copy · Models: export · Models: list · Models: list checkpoints 4. ### Featurestores entity types 14 6 read · 6 write · 2 destructive Featurestores entity types: create · Featurestores entity types: export feature values · Featurestores entity types: import feature values · Featurestores entity types: list 5. ### Feature online stores feature views 13 6 read · 6 write · 1 destructive Feature online stores feature views: create · Feature online stores feature views: direct write · Feature online stores feature views: fetch feature values · Feature online stores feature views: generate fetch access token 6. ### Reasoning engines 12 4 read · 6 write · 2 destructive Reasoning engines: async query · Reasoning engines: create · Reasoning engines: execute code · Reasoning engines: list 7. ### Datasets 11 5 read · 5 write · 1 destructive Datasets: list · Datasets: create · Datasets: export · Datasets: import 8. ### Featurestores 10 5 read · 4 write · 1 destructive Featurestores: batch read feature values · Featurestores: create · Featurestores: list · Featurestores: search features **112 more areas · 1,014 actions** - Index endpoints **10** - Metadata stores contexts **10** - Publishers models **10** - Reasoning engines memories **10** - Studies trials **10** - Express reasoning engines **9** - Express reasoning engines sandbox environments **9** - Memory banks memories **9** - Reasoning engines sandbox environments **9** - Agents **8** - Express publishers models **8** - Feature groups **8** - Feature online stores **8** - Metadata stores executions **8** - Model deployment monitoring jobs **8** - Notebook runtime templates **8** - Tensorboards **8** - Tensorboards experiments runs time series **8** - Express endpoints **7** - Indexes **7** - Metadata stores artifacts **7** - Notebook runtimes **7** - Online evaluators **7** - Pipeline jobs **7** - Reasoning engines sessions **7** - Schedules **7** - Tensorboards experiments **7** - Tensorboards experiments runs **7** - Datasets dataset versions **6** - Deployment resource pools **6** - Evaluation sets **6** - Express datasets dataset versions **6** - Feature groups features **6** - Featurestores entity types features **6** - Persistent resources **6** - Batch prediction jobs **5** - Cached contents **5** - Custom jobs **5** - Custom jobs operations **5** - Data labeling jobs **5** - Data labeling jobs operations **5** - Datasets annotation specs operations **5** - Datasets data items annotations operations **5** - Datasets data items operations **5** - Datasets operations **5** - Datasets saved queries operations **5** - Deployment resource pools operations **5** - Endpoints operations **5** - Evaluation runs **5** - Express custom jobs operations **5** - Express data labeling jobs operations **5** - Express datasets **5** - Express datasets annotation specs operations **5** - Express datasets data items annotations operations **5** - Express datasets data items operations **5** - Express datasets operations **5** - Express datasets saved queries operations **5** - Express deployment resource pools operations **5** - Express endpoints operations **5** - Express featurestores entity types features operations **5** - Express featurestores entity types operations **5** - Express featurestores operations **5** - Express hyperparameter tuning jobs operations **5** - Express index endpoints operations **5** - Express indexes operations **5** - Express memory banks memories operations **5** - Express memory banks operations **5** - Express metadata stores artifacts operations **5** - Express metadata stores contexts operations **5** - Express metadata stores executions operations **5** - Express metadata stores operations **5** - Express migratable resources operations **5** - Express model deployment monitoring jobs operations **5** - Express models evaluations operations **5** - Express models operations **5** - Express notebook execution jobs operations **5** - Express notebook runtime templates operations **5** - Express notebook runtimes operations **5** - Express online evaluators operations **5** - Express operations **5** - Express persistent resources operations **5** - Express pipeline jobs operations **5** - Express rag corpora operations **5** - Express rag corpora rag files operations **5** - Express rag engine config operations **5** - Express reasoning engines memories operations **5** - Express reasoning engines operations **5** - Express reasoning engines sessions operations **5** - Express schedules operations **5** - Express semantic governance policies operations **5** - Express serving profiles operations **5** - Express skills operations **5** - Express specialist pools operations **5** - Express studies operations **5** - Express studies trials operations **5** - Express task stores operations **5** - Express tensorboards experiments operations **5** - Express tensorboards experiments runs operations **5** - Express tensorboards experiments runs time series operations **5** - Express tensorboards operations **5** - Express training pipelines operations **5** - Express tuning jobs operations **5** - Featurestores entity types features operations **5** - Featurestores entity types operations **5** - Featurestores operations **5** - Hyperparameter tuning jobs **5** - Hyperparameter tuning jobs operations **5** - Index endpoints operations **5** - Indexes operations **5** - Memory banks memories operations **5** - Memory banks operations **5** - 99 smaller areas **364** Top actions ## The Google Vertex AI API, shaped for agents. The ids your agent types, what each does, and the one name it shares with other systems. 9 of Google Vertex AI's 9 action names are shared with at least one other connector, so an agent learns them once. | Action | What it does | Effect | Same name in | | --- | --- | --- | --- | | `gcp_vertex_ai.endpoints.predict` Endpoints: predict | Endpoints: predict. | write | `analytics.inference.create` 17 connectors | | `gcp_vertex_ai.endpoints.raw_predict` Endpoints: raw predict | Perform an online prediction with an arbitrary HTTP payload. | write | `analytics.inference.create` 17 connectors | | `gcp_vertex_ai.batch_prediction_jobs.list` Batch prediction jobs: list | Lists BatchPredictionJobs in a Location. | read | `automation.background_job.list` 34 connectors | | `gcp_vertex_ai.batch_prediction_jobs.create` Batch prediction jobs: create | Creates a BatchPredictionJob. | write | `automation.background_job.create` 29 connectors | | `gcp_vertex_ai.locations.list` Locations: list | Lists information about the supported locations for this service. | read | `infra.region.list` 16 connectors | | `gcp_vertex_ai.batch_prediction_jobs.cancel` Batch prediction jobs: cancel | Cancels a BatchPredictionJob. | destructive | `automation.background_job.cancel` 17 connectors | | `gcp_vertex_ai.datasets.list` Datasets: list | Lists Datasets in a Location. | read | this connector | | `gcp_vertex_ai.custom_jobs.list` Custom jobs: list | Lists CustomJobs in a Location. | read | this connector | | `gcp_vertex_ai.locations.get` Locations: get | Gets information about a location. | read | `infra.region.get` 4 connectors | | `gcp_vertex_ai.express_datasets.list` Express datasets: list | Lists Datasets in a Location. | read | this connector | | `gcp_vertex_ai.express_datasets.create` Express datasets: create | Creates a Dataset. | write | this connector | | `gcp_vertex_ai.express_endpoints.predict` Express endpoints: predict | Express endpoints: predict. | write | this connector | Try it ## Prompts to paste into Claude. Each one names the Google Vertex AI actions it will call, so you know what it reads and what it changes before you send it. 1. Read read Express batch prediction jobs: list in Google Vertex AI and tell me what stands out. calls `gcp_vertex_ai.express_batch_prediction_jobs.list` 2. Write write Endpoints: predict in Google Vertex AI. Show me exactly what you will send before you save it. calls `gcp_vertex_ai.endpoints.predict` Governance for Google Vertex AI ## Destructive actions wait for a yes. 321 Google Vertex AI actions are destructive. The first 10: - Batch prediction jobs: cancel `gcp_vertex_ai.batch_prediction_jobs.cancel` - Batch prediction jobs: delete `gcp_vertex_ai.batch_prediction_jobs.delete` - Models: delete `gcp_vertex_ai.models.delete` - Indexes: delete `gcp_vertex_ai.indexes.delete` - Datasets: delete `gcp_vertex_ai.datasets.delete` - Endpoints: delete `gcp_vertex_ai.endpoints.delete` - Operations: delete `gcp_vertex_ai.operations.delete` - Custom jobs: delete `gcp_vertex_ai.custom_jobs.delete` - Rag corpora: delete `gcp_vertex_ai.rag_corpora.delete` - Featurestores: delete `gcp_vertex_ai.featurestores.delete` A destructive action runs only with `confirm: true`. Under the `human` policy the agent cannot confirm for itself: the run stops until a named person approves, in the console or from a signed link in an email. The cap on a key is refused before the call is sent, never after. do gcp_vertex_ai.batch_prediction_jobs.cancel confirm_policy: **human** Batch prediction jobs: cancel? **intent:** the user asked me to batch prediction jobs: cancel run suspended · waiting for a named person · key _gcp-vertex-ai-agent_ cannot confirm itself In the console, or from an emailed signed link. a key for this connector ``` account { op: "keys.create", name: "gcp-vertex-ai-agent", allowed_actions: ["gcp_vertex_ai.*"], cap_calls: 2000, confirm_policy: "human_for_writes", approval_rules: [ { match: "gcp_vertex_ai.batch_prediction_jobs.cancel", approver: "lead@example.com" }, { match: "gcp_vertex_ai.batch_prediction_jobs.delete", approver: "lead@example.com" }, { match: "gcp_vertex_ai.models.delete", approver: "lead@example.com" } ] } ``` Illustrative. Scoped to Google Vertex AI only, capped at 2,000 upstream calls, every write confirmed, and the destructive ones approved by a named person. Cost ## What Google Vertex AI costs on actionlake. Every Google Vertex AI action is one upstream call, the only thing actionlake bills. Searching, `check`, `cost`, `account` and any call Google Vertex AI refused with 401, 403, 429 or a 5xx are free. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000. - 1 upstream call per action all 1,128 Google Vertex AI actions - 196 actions page each page of a long list is one call - 217 reads can be cached a cache hit costs nothing - 100,000 calls a month, free paid plans from US$49 a month for 250,000 [See pricing](https://actionlake.com/pricing) Action index ## Every Google Vertex AI action. 1,128 actions by area: the id your agent calls, what it does, and its effect. **Endpoints 21** - `gcp_vertex_ai.endpoints.compute_tokens` Endpoints: compute tokens read - `gcp_vertex_ai.endpoints.generate_content` Endpoints: generate content write - and 19 more in this area **Locations 18** - `gcp_vertex_ai.locations.ask_contexts` Locations: ask contexts write - `gcp_vertex_ai.locations.async_retrieve_contexts` Locations: async retrieve contexts write - and 16 more in this area **Models 15** - `gcp_vertex_ai.models.copy` Models: copy write - `gcp_vertex_ai.models.export` Models: export write - and 13 more in this area **Featurestores entity types 14** - `gcp_vertex_ai.featurestores_entity_types.create` Featurestores entity types: create write - `gcp_vertex_ai.featurestores_entity_types.export_feature_values` Featurestores entity types: export feature values write - and 12 more in this area **Feature online stores feature views 13** - `gcp_vertex_ai.feature_online_stores_feature_views.create` Feature online stores feature views: create write - `gcp_vertex_ai.feature_online_stores_feature_views.direct_write` Feature online stores feature views: direct write write - and 11 more in this area **Reasoning engines 12** - `gcp_vertex_ai.reasoning_engines.async_query` Reasoning engines: async query write - `gcp_vertex_ai.reasoning_engines.create` Reasoning engines: create write - and 10 more in this area **Datasets 11** - `gcp_vertex_ai.datasets.create` Datasets: create write - `gcp_vertex_ai.datasets.export` Datasets: export write - and 9 more in this area **Featurestores 10** - `gcp_vertex_ai.featurestores.batch_read_feature_values` Featurestores: batch read feature values write - `gcp_vertex_ai.featurestores.create` Featurestores: create write - and 8 more in this area **Index endpoints 10** - `gcp_vertex_ai.index_endpoints.create` Index endpoints: create write - `gcp_vertex_ai.index_endpoints.deploy_index` Index endpoints: deploy index write - and 8 more in this area **Metadata stores contexts 10** - `gcp_vertex_ai.metadata_stores_contexts.add_context_artifacts_and_executions` Metadata stores contexts: add context artifacts and executions write - `gcp_vertex_ai.metadata_stores_contexts.add_context_children` Metadata stores contexts: add context children write - and 8 more in this area **Publishers models 10** - `gcp_vertex_ai.publishers_models.compute_tokens` Publishers models: compute tokens read - `gcp_vertex_ai.publishers_models.count_tokens` Publishers models: count tokens read - and 8 more in this area **Reasoning engines memories 10** - `gcp_vertex_ai.reasoning_engines_memories.create` Reasoning engines memories: create write - `gcp_vertex_ai.reasoning_engines_memories.generate` Reasoning engines memories: generate write - and 8 more in this area **Studies trials 10** - `gcp_vertex_ai.studies_trials.add_trial_measurement` Studies trials: add trial measurement write - `gcp_vertex_ai.studies_trials.check_trial_early_stopping_state` Studies trials: check trial early stopping state write - and 8 more in this area **Express reasoning engines 9** - `gcp_vertex_ai.express_reasoning_engines.async_query` Express reasoning engines: async query write - `gcp_vertex_ai.express_reasoning_engines.create` Express reasoning engines: create write - and 7 more in this area **Express reasoning engines sandbox environments 9** - `gcp_vertex_ai.express_reasoning_engines_sandbox_environments.authorize_access` Express reasoning engines sandbox environments: authorize access write - `gcp_vertex_ai.express_reasoning_engines_sandbox_environments.create` Express reasoning engines sandbox environments: create write - and 7 more in this area **Memory banks memories 9** - `gcp_vertex_ai.memory_banks_memories.create` Memory banks memories: create write - `gcp_vertex_ai.memory_banks_memories.generate` Memory banks memories: generate write - and 7 more in this area **Reasoning engines sandbox environments 9** - `gcp_vertex_ai.reasoning_engines_sandbox_environments.authorize_access` Reasoning engines sandbox environments: authorize access write - `gcp_vertex_ai.reasoning_engines_sandbox_environments.create` Reasoning engines sandbox environments: create write - and 7 more in this area **Agents 8** - `gcp_vertex_ai.agents.create` Agents: create write - `gcp_vertex_ai.agents.get` Agents: get read - and 6 more in this area **Express publishers models 8** - `gcp_vertex_ai.express_publishers_models.compute_tokens` Express publishers models: compute tokens read - `gcp_vertex_ai.express_publishers_models.count_tokens` Express publishers models: count tokens read - and 6 more in this area **Feature groups 8** - `gcp_vertex_ai.feature_groups.create` Feature groups: create write - `gcp_vertex_ai.feature_groups.get` Feature groups: get read - and 6 more in this area **Feature online stores 8** - `gcp_vertex_ai.feature_online_stores.create` Feature online stores: create write - `gcp_vertex_ai.feature_online_stores.get` Feature online stores: get read - and 6 more in this area **Metadata stores executions 8** - `gcp_vertex_ai.metadata_stores_executions.add_execution_events` Metadata stores executions: add execution events write - `gcp_vertex_ai.metadata_stores_executions.create` Metadata stores executions: create write - and 6 more in this area **Model deployment monitoring jobs 8** - `gcp_vertex_ai.model_deployment_monitoring_jobs.create` Model deployment monitoring jobs: create write - `gcp_vertex_ai.model_deployment_monitoring_jobs.get` Model deployment monitoring jobs: get read - and 6 more in this area **Notebook runtime templates 8** - `gcp_vertex_ai.notebook_runtime_templates.create` Notebook runtime templates: create write - `gcp_vertex_ai.notebook_runtime_templates.get` Notebook runtime templates: get read - and 6 more in this area **Tensorboards 8** - `gcp_vertex_ai.tensorboards.batch_read` Tensorboards: batch read read - `gcp_vertex_ai.tensorboards.create` Tensorboards: create write - and 6 more in this area **Tensorboards experiments runs time series 8** - `gcp_vertex_ai.tensorboards_experiments_runs_time_series.create` Tensorboards experiments runs time series: create write - `gcp_vertex_ai.tensorboards_experiments_runs_time_series.export_tensorboard_time_series` Tensorboards experiments runs time series: export tensorboard time series read - and 6 more in this area **Express endpoints 7** - `gcp_vertex_ai.express_endpoints.compute_tokens` Express endpoints: compute tokens read - `gcp_vertex_ai.express_endpoints.count_tokens` Express endpoints: count tokens read - and 5 more in this area **Indexes 7** - `gcp_vertex_ai.indexes.create` Indexes: create write - `gcp_vertex_ai.indexes.get` Indexes: get read - and 5 more in this area **Metadata stores artifacts 7** - `gcp_vertex_ai.metadata_stores_artifacts.create` Metadata stores artifacts: create write - `gcp_vertex_ai.metadata_stores_artifacts.get` Metadata stores artifacts: get read - and 5 more in this area **Notebook runtimes 7** - `gcp_vertex_ai.notebook_runtimes.assign` Notebook runtimes: assign write - `gcp_vertex_ai.notebook_runtimes.get` Notebook runtimes: get read - and 5 more in this area Online evaluators **7** Pipeline jobs **7** Reasoning engines sessions **7** Schedules **7** Tensorboards experiments **7** Tensorboards experiments runs **7** Datasets dataset versions **6** Deployment resource pools **6** Evaluation sets **6** Express datasets dataset versions **6** Feature groups features **6** Featurestores entity types features **6** Persistent resources **6** Batch prediction jobs **5** Cached contents **5** Custom jobs **5** Custom jobs operations **5** Data labeling jobs **5** Data labeling jobs operations **5** Datasets annotation specs operations **5** Datasets data items annotations operations **5** Datasets data items operations **5** Datasets operations **5** Datasets saved queries operations **5** Deployment resource pools operations **5** Endpoints operations **5** Evaluation runs **5** Express custom jobs operations **5** Express data labeling jobs operations **5** Express datasets **5** Express datasets annotation specs operations **5** Express datasets data items annotations operations **5** Express datasets data items operations **5** Express datasets operations **5** Express datasets saved queries operations **5** Express deployment resource pools operations **5** Express endpoints operations **5** Express featurestores entity types features operations **5** Express featurestores entity types operations **5** Express featurestores operations **5** Express hyperparameter tuning jobs operations **5** Express index endpoints operations **5** Express indexes operations **5** Express memory banks memories operations **5** Express memory banks operations **5** Express metadata stores artifacts operations **5** Express metadata stores contexts operations **5** Express metadata stores executions operations **5** Express metadata stores operations **5** Express migratable resources operations **5** Express model deployment monitoring jobs operations **5** Express models evaluations operations **5** Express models operations **5** Express notebook execution jobs operations **5** Express notebook runtime templates operations **5** Express notebook runtimes operations **5** Express online evaluators operations **5** Express operations **5** Express persistent resources operations **5** Express pipeline jobs operations **5** Express rag corpora operations **5** Express rag corpora rag files operations **5** Express rag engine config operations **5** Express reasoning engines memories operations **5** Express reasoning engines operations **5** Express reasoning engines sessions operations **5** Express schedules operations **5** Express semantic governance policies operations **5** Express serving profiles operations **5** Express skills operations **5** Express specialist pools operations **5** Express studies operations **5** Express studies trials operations **5** Express task stores operations **5** Express tensorboards experiments operations **5** Express tensorboards experiments runs operations **5** Express tensorboards experiments runs time series operations **5** Express tensorboards operations **5** Express training pipelines operations **5** Express tuning jobs operations **5** Featurestores entity types features operations **5** Featurestores entity types operations **5** Featurestores operations **5** Hyperparameter tuning jobs **5** Hyperparameter tuning jobs operations **5** Index endpoints operations **5** Indexes operations **5** Memory banks memories operations **5** Memory banks operations **5** 99 smaller areas **364** This index shows a sample of each area. All 1,128 actions, with their inputs and prices, are in the [catalogue](https://actionlake.com/console/catalogue), which a free sandbox key can search. Questions ## The Google Vertex AI MCP server, answered. **Is there a Google Vertex AI MCP server for Claude?** Yes. actionlake is an MCP gateway: a remote MCP server that gives Claude, or another MCP client, 1,128 Google Vertex AI actions across endpoints, locations, models and featurestores entity types, through the same eight tools it uses for every other connector. actionlake is independent; it is not made or endorsed by Google Vertex AI. **What do I need to connect Google Vertex AI?** Your own Google Vertex AI OAuth app, registered with Google Vertex AI (its client id and client secret), and the access and refresh tokens from one admin sign-in. actionlake has no Google Vertex AI app of its own, so this is not a one-step consent. actionlake renews the access token as it expires. Plus your host. The free call account { op: "connections.requirements", service: "gcp_vertex_ai" } lists exactly what to paste. **Can an AI agent delete or change things in Google Vertex AI?** Only if its key allows it, and only with an explicit confirmation. 321 Google Vertex AI actions are destructive, including batch prediction jobs: cancel, batch prediction jobs: delete and models: delete. Each runs only with confirm: true. Under the human confirm policy, or an approval rule on the key, the run waits until a named person approves it. **Will an AI agent use up my Google Vertex AI API limit?** actionlake paces every call to the limits in the Google Vertex AI service file: 600 calls a minute per connection. Repeated reads are answered from cache (217 Google Vertex AI actions can be), which never reaches Google Vertex AI. **What does a Google Vertex AI action cost?** One upstream call, the only unit actionlake bills. A list that pages bills one call per page (196 Google Vertex AI actions page). Cache hits, searches, dry runs and calls Google Vertex AI refused with 401, 403, 429 or a 5xx are free. The Free plan includes 100,000 calls a month with no card. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000. **Which Google Vertex AI API does it cover?** The Google Vertex AI (Gemini generation, model endpoints, model registry, datasets, training, pipelines, RAG, Agent Engine) API aiplatform v1 (Discovery revision 20261012) at aiplatform.googleapis.com by default, or the base URL set when you connect. It is built from Google Vertex AI's documentation as read on 22 October 2026. 1,128 actions in 218 areas, each one in the signed service file (version 1.0.0). **Does actionlake keep my Google Vertex AI data, and where?** actionlake processes and stores data in Sydney, Australia; there is no US region yet. Reads may be cached for up to 5 minutes, and a write clears the reads it affects. Each run's request and response is kept, sealed, for your plan's trace window: 1 day on Free, 3 days on Starter, 7 days on Pro, 14 days on Team, 30 days on Business and 30 days on Enterprise. Your Google Vertex AI credential is sealed under your tenant's key, never logged and never returned. Your traffic is not used to train models. **Which AI clients can use it?** Claude Code, tested by us end to end, and the Claude app as a custom connector, tested by us with an existing key. Other MCP clients that support remote servers over Streamable HTTP (MCP revisions 2025-06-18 and 2025-11-25) and OAuth dynamic client registration, or a bearer header, should connect the same way; we have not tested them. Or plain REST, under the same action names. ## More connectors like Google Vertex AI Shares the most action names with - [Google Gemini API (AI Studio) 7](https://actionlake.com/connectors/google-gemini-api-ai-studio) - [Claude API (Anthropic) 7](https://actionlake.com/connectors/anthropic-claude-api) - [Perplexity 6](https://actionlake.com/connectors/perplexity) - [Together AI 6](https://actionlake.com/connectors/together-ai) - [Groq 5](https://actionlake.com/connectors/groq) - [xAI Grok API 5](https://actionlake.com/connectors/xai-grok-api) Joins with, on shared entities - [NetSuite 1](https://actionlake.com/connectors/netsuite) - [Microsoft Fabric 1](https://actionlake.com/connectors/microsoft-fabric) - [Kubernetes 1](https://actionlake.com/connectors/kubernetes) - [Tableau 1](https://actionlake.com/connectors/tableau) - [Veeam Backup & Replication 1](https://actionlake.com/connectors/veeam-backup-and-replication) - [Veeam Backup for Microsoft 365 1](https://actionlake.com/connectors/veeam-backup-for-microsoft-365) Browse - [All AI connectors](https://actionlake.com/connectors/category/ai) - [All connectors](https://actionlake.com/connectors) - [How actionlake compares](https://actionlake.com/compare) actionlake is independent. It is not made or endorsed by Google Vertex AI; the name is the vendor's own. Counts are from the Google Vertex AI service file, version 1.0.0. ## Give your agent Google Vertex AI today. [Connect Google Vertex AI free](https://actionlake.com/console?start=sandbox) [See pricing](https://actionlake.com/pricing) Start free - 100,000 calls/month, no card. No signup. Hard caps on every key. Cached answers are free. --- Source: https://actionlake.com/connectors/google-vertex-ai More: https://actionlake.com/llms.txt