--- title: "Google Cloud Document AI MCP server - actionlake" description: "Give Claude 42 Google Cloud Document AI actions through one MCP server: operations. Your agent never holds the credential. 100,000 calls a month free." canonical: https://actionlake.com/connectors/google-cloud-document-ai updated: 2026-10-23 --- Connectors · AI # Google Cloud Document AI MCP server for AI agents. actionlake is an MCP gateway: one remote MCP server that connects Claude and other MCP clients to Google Cloud Document AI and the rest of your systems. It gives your agent 42 Google Cloud Document AI actions (18 reads, 18 writes and 6 that wait for a confirmation) and holds the Google Cloud Document AI credential itself, so the agent never does. It paces Google Cloud Document AI's limits and records every call. [Connect Google Cloud Document 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: "google_document_ai.operations.list", intent: "check which Google Cloud Document AI account is connected" } # Lists operations that match the specified filter in the request · read · 1 upstream call ``` - 42 actions in one signed file - 18 read look, never change - 18 write create and update - 6 wait for a yes destructive, confirmed first Set up ## Connect Google Cloud Document 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 Cloud Document AI. Signed token (JWT). You paste the signing key once; actionlake signs a short-lived token for each call. actionlake stores what you paste write-only, sealed under your tenant's key. - `client_email` Service account client email (`client_email` in the JSON key file) - `private_key` Service account private key, PKCS#8 PEM (`private_key` in the JSON key file) - `host` Where your Google Cloud Document AI lives: your host - `location` Your Google Cloud Document AI location - `project_id` Your Google Cloud Document AI project id ``` account { op: "connections.requirements", service: "google_document_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: "google_document_ai.operations.list" } # lists operations that match the specified filter in the request · read · 1 upstream call check { action: "google_document_ai.processors.create", input: { … } } # free: inputs, permissions, cost and confirmations, nothing sent ``` Google Cloud Document AI at a glance **Sign-in** Signed token (JWT). You paste the signing key once; actionlake signs a short-lived token for each call. **You provide** `client_email` `private_key` **Setup fields** `host` `location` `project_id` **Limits we pace to** 5 calls a second, per connection **API covered** Google Cloud Document AI API v1, at your Google Cloud Document AI instance's API base URL, which you enter when connecting **Built from** Google Cloud Document AI's documentation, read 23 October 2026[documentai.googleapis.com](https://documentai.googleapis.com/$discovery/rest?version=v1) [cloud.google.com](https://cloud.google.com/document-ai/docs/reference/rest) **Service file** Version 1.0.0, signed (Ed25519) **Cached reads** 18 of 18 reads, for up to an hour; a write clears what it affects **Shared entities** 2, so ask can line Google Cloud Document 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 Cloud Document AI. 42 actions in 1 area. Every one carries its effect, so a key can allow the reads and hold the rest. - **18** read look, never change - **18** write create and update - **6** destructive wait for a confirmation 1. ### Document processing 42 18 read · 18 write · 6 destructive Lists operations that match the specified filter in the request · Lists all processors which belong to this project · Lists the processor types that exist · LRO endpoint to batch process many documents Top actions ## The Google Cloud Document AI API, shaped for agents. The ids your agent types, what each does, and the one name it shares with other systems. | Action | What it does | Effect | Same name in | | --- | --- | --- | --- | | `google_document_ai.operations.list` Lists operations that match the specified filter in the request | Lists operations that match the specified filter in the request. | read | this connector | | `google_document_ai.processors.list` Lists all processors which belong to this project | Lists all processors which belong to this project. | read | this connector | | `google_document_ai.processors.create` Creates a processor from the ProcessorType provided | Creates a processor from the ProcessorType provided. | write | this connector | | `google_document_ai.processor_types.list` Lists the processor types that exist | Lists the processor types that exist. | read | this connector | | `google_document_ai.operations.get` Gets the latest state of a long-running operation | Gets the latest state of a long-running operation. | read | this connector | | `google_document_ai.processor_types.get` Gets a processor type detail | Gets a processor type detail. | read | this connector | | `google_document_ai.processors.delete` delete processors | Deletes the processor, unloads all deployed model artifacts if it was enabled and then deletes all artifacts associated with this processor. | destructive | this connector | | `google_document_ai.schemas.list` Lists Schemas | Lists Schemas. | read | this connector | | `google_document_ai.locations.list` Lists information about the supported locations for this service | Lists information about the supported locations for this service. | read | this connector | | `google_document_ai.schemas.create` Creates a schema | Creates a schema. | write | this connector | | `google_document_ai.evaluations.list` Retrieves a set of evaluations for a given processor version | Retrieves a set of evaluations for a given processor version. | read | this connector | | `google_document_ai.schema_versions.create` Creates a schema version | Creates a schema version. | write | this connector | Try it ## Prompts to paste into Claude. Each one names the Google Cloud Document AI actions it will call, so you know what it reads and what it changes before you send it. 1. Read read Lists operations that match the specified filter in the request in Google Cloud Document AI and tell me what stands out. calls `google_document_ai.operations.list` 2. Write write Creates a processor from the ProcessorType provided in Google Cloud Document AI. Show me exactly what you will send before you save it. calls `google_document_ai.processors.create` Governance for Google Cloud Document AI ## Destructive actions wait for a yes. 6 Google Cloud Document AI actions are destructive. These are all of them: - delete processors `google_document_ai.processors.delete` - Starts asynchronous cancellation on a long-running operation `google_document_ai.operations.cancel` - Deletes a schema `google_document_ai.schemas.delete` - Deletes a schema version `google_document_ai.schema_versions.delete` - Deletes a long-running operation `google_document_ai.global_operations.delete` - Deletes the processor version, all artifacts under the processor version will be deleted `google_document_ai.processor_versions.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 google_document_ai.processors.delete confirm_policy: **human** delete processors? **intent:** the user asked me to delete processors run suspended · waiting for a named person · key _google-document-ai-agent_ cannot confirm itself In the console, or from an emailed signed link. a key for this connector ``` account { op: "keys.create", name: "google-document-ai-agent", allowed_actions: ["google_document_ai.*"], cap_calls: 2000, confirm_policy: "human_for_writes", approval_rules: [ { match: "google_document_ai.processors.delete", approver: "lead@example.com" }, { match: "google_document_ai.operations.cancel", approver: "lead@example.com" }, { match: "google_document_ai.schemas.delete", approver: "lead@example.com" } ] } ``` Illustrative. Scoped to Google Cloud Document AI only, capped at 2,000 upstream calls, every write confirmed, and the destructive ones approved by a named person. Cost ## What Google Cloud Document AI costs on actionlake. Every Google Cloud Document AI action is one upstream call, the only thing actionlake bills. Searching, `check`, `cost`, `account` and any call Google Cloud Document 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 42 Google Cloud Document AI actions - 42 actions page each page of a long list is one call - 18 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 Cloud Document AI action. 42 actions by area: the id your agent calls, what it does, and its effect. **Document processing 42** - `google_document_ai.evaluations.get` Retrieves a specific evaluation read - `google_document_ai.evaluations.list` Retrieves a set of evaluations for a given processor version read - `google_document_ai.global_operations.delete` Deletes a long-running operation destructive - `google_document_ai.human_review.review_document` Send a document for Human Review write - `google_document_ai.locations.fetch_processor_types` Fetches processor types read - `google_document_ai.locations.get` Gets information about a location read - `google_document_ai.locations.list` Lists information about the supported locations for this service read - `google_document_ai.operations.cancel` Starts asynchronous cancellation on a long-running operation destructive - `google_document_ai.operations.get` Gets the latest state of a long-running operation read - `google_document_ai.operations.list` Lists operations that match the specified filter in the request read - `google_document_ai.processor_types.get` Gets a processor type detail read - `google_document_ai.processor_types.list` Lists the processor types that exist read - `google_document_ai.processor_versions.batch_process` LRO endpoint to batch process many documents write - `google_document_ai.processor_versions.delete` Deletes the processor version, all artifacts under the processor version will be deleted destructive - `google_document_ai.processor_versions.deploy` Deploys the processor version write - `google_document_ai.processor_versions.evaluate_processor_version` Evaluates a ProcessorVersion against annotated documents, producing an Evaluation write - `google_document_ai.processor_versions.get` Gets a processor version detail read - `google_document_ai.processor_versions.list` Lists all versions of a processor read - `google_document_ai.processor_versions.process` Processes a single document write - `google_document_ai.processor_versions.train` Trains a new processor version write - `google_document_ai.processor_versions.undeploy` Undeploys the processor version write - `google_document_ai.processors.batch_process` LRO endpoint to batch process many documents write - `google_document_ai.processors.create` Creates a processor from the ProcessorType provided write - `google_document_ai.processors.delete` delete processors destructive - `google_document_ai.processors.disable` Disables a processor write - `google_document_ai.processors.enable` Enables a processor write - `google_document_ai.processors.get` Gets a processor detail read - `google_document_ai.processors.list` Lists all processors which belong to this project read - `google_document_ai.processors.process` Processes a single document write - `google_document_ai.processors.set_default_processor_version` setDefaultProcessorVersion processors write - `google_document_ai.project_operations.get` Gets the latest state of a long-running operation read - `google_document_ai.schema_versions.create` Creates a schema version write - `google_document_ai.schema_versions.delete` Deletes a schema version destructive - `google_document_ai.schema_versions.generate` Generates a schema version write - `google_document_ai.schema_versions.get` Gets a schema version read - `google_document_ai.schema_versions.list` Lists SchemaVersions read - `google_document_ai.schema_versions.patch` Updates a schema version write - `google_document_ai.schemas.create` Creates a schema write - `google_document_ai.schemas.delete` Deletes a schema destructive - `google_document_ai.schemas.get` Gets a schema read - `google_document_ai.schemas.list` Lists Schemas read - `google_document_ai.schemas.patch` Updates a schema write Questions ## The Google Cloud Document AI MCP server, answered. **Is there a Google Cloud Document AI MCP server for Claude?** Yes. actionlake is an MCP gateway: a remote MCP server that gives Claude, or another MCP client, 42 Google Cloud Document AI actions, through the same eight tools it uses for every other connector. actionlake is independent; it is not made or endorsed by Google Cloud Document AI. **What do I need to connect Google Cloud Document AI?** Service account client email (`client_email` in the JSON key file) and Service account private key, PKCS#8 PEM (`private_key` in the JSON key file). Plus your host, your location and your project id. The free call account { op: "connections.requirements", service: "google_document_ai" } lists exactly what to paste. **Can an AI agent delete or change things in Google Cloud Document AI?** Only if its key allows it, and only with an explicit confirmation. 6 Google Cloud Document AI actions are destructive, including delete processors, starts asynchronous cancellation on a long-running operation and deletes a schema. 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 Cloud Document AI API limit?** actionlake paces every call to the limits in the Google Cloud Document AI service file: 5 calls a second per connection. Repeated reads are answered from cache (18 Google Cloud Document AI actions can be), which never reaches Google Cloud Document AI. **What does a Google Cloud Document AI action cost?** One upstream call, the only unit actionlake bills. A list that pages bills one call per page (42 Google Cloud Document AI actions page). Cache hits, searches, dry runs and calls Google Cloud Document 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 Cloud Document AI API does it cover?** The Google Cloud Document AI API v1 at your Google Cloud Document AI instance's API base URL, which you enter when connecting. It is built from Google Cloud Document AI's documentation as read on 23 October 2026. 42 actions in 1 area, each one in the signed service file (version 1.0.0). **Does actionlake keep my Google Cloud Document 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 an hour, 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 Cloud Document 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 Cloud Document AI Joins with, on shared entities - [Kubernetes 2](https://actionlake.com/connectors/kubernetes) - [Northflank 2](https://actionlake.com/connectors/northflank) - [Azure Container Apps 2](https://actionlake.com/connectors/azure-container-apps) - [Google Cloud Run 2](https://actionlake.com/connectors/google-cloud-run) - [DigitalOcean 1](https://actionlake.com/connectors/digitalocean) - [Azure App Service (Web Apps and Functions) 1](https://actionlake.com/connectors/azure-app-service-web-apps-and-functions) 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 Cloud Document AI; the name is the vendor's own. Counts are from the Google Cloud Document AI service file, version 1.0.0. ## Give your agent Google Cloud Document AI today. [Connect Google Cloud Document 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-cloud-document-ai More: https://actionlake.com/llms.txt