Google Contacts MCP Tools for AI Agents
Let an agent search contacts by name, email, or phone, look up your organization directory, and create contact records in Google Contacts.
Available Tools
8 tools available in this integration
GoogleContacts.CreateContact Create a new contact record in Google Contacts.
GoogleContacts.SearchContacts Search the user's contacts in Google Contacts by name, email address or phone number.
GoogleContacts.SearchContactsByEmail Search the user's contacts in Google Contacts by email address.
GoogleContacts.SearchContactsByName Search the user's contacts in Google Contacts by name.
GoogleContacts.SearchContactsByPhoneNumber Search the user's contacts in Google Contacts by phone number.
GoogleContacts.SearchDirectoryPeople Search the user's Google Workspace organization directory by name or email.
GoogleContacts.SearchOtherContacts Search the user's other contacts: people Google saved automatically because the user
GoogleContacts.WhoAmI Get comprehensive user profile and Google Contacts environment information.
Why agent-optimized tools matter
Built for intent, not API schemas
Ask the agent to "find the person behind this phone number and add them if they are not already in my contacts." Arcade picks the right Google Contacts actions without guessing which search method to use first.
Up to 100x fewer tokens
Google Contacts searches return only matching contact or directory records by default, so the agent can identify the right person without loading your entire address book or Workspace directory into the model.
Complete the workflow end to end
Search your contacts, check automatically saved people, look up a coworker in the directory, create the missing contact, and return what changed through one toolkit designed around the complete contact task.
Make actions predictable and controlled
Arcade tools return structured results and useful errors across users and runs. Teams can apply scope restrictions, redaction, and approval policies so agents follow the same controls every time.
Securing Google Contacts
The more capabilities agents get, the more useful they are. However, more capabilities bring more risks of undesired effects such as data loss and exfiltration.
This is why Arcade provides you with contextual rules that dictate how and when a tool is allowed to run. Arcade’s contextual access lets you govern tool visibility and behavior at every stage of execution and control who can see a tool, what inputs are allowed, and what comes back.
Learn about Contextual Access PoliciesFor Google Contacts, you can implement some of these policies:
- Hide GoogleContacts.CreateContact from agents unless the roles your IDP marks as contact managers are active, so regular agents cannot add unreviewed people to users' address books.
- Require human approval before GoogleContacts.CreateContact runs, so an agent cannot create fraudulent or mistaken contact records.
- Strip phone numbers and postal addresses from GoogleContacts.SearchContactsByEmail results, so matching a person by email does not expose unrelated personal details.
- Redact personal phone numbers and home addresses from GoogleContacts.SearchDirectoryPeople results, so the agent receives directory identities without private contact data.
- Redact account identifiers, recovery contact details, and organization-specific environment identifiers from GoogleContacts.WhoAmI results before they reach the model.
How to use Google Contacts with your agent
Connect Google Contacts to an MCP client and give your agent access to the contact workflows it needs. Follow the walkthrough to configure authorization and run your first prompt.
Your questions, answered
A Google Contacts MCP server exposes contact and directory capabilities as structured tools an AI agent can call through the Model Context Protocol. After a user authorizes their Google account, the agent can search contacts, look up directory people, inspect automatically saved contacts, and create contact records.
Google Contacts does not provide the first-party MCP endpoint used here. Arcade provides the MCP runtime around Google Contacts, including per-user OAuth, approval workflows, redaction, scope controls, and access from supported agent frameworks and clients.
Arcade agents can search contacts by name, email address, or phone number; search a Google Workspace directory; find people Google saved automatically; identify the authorized user's Contacts environment; and create new contact records. They cannot update, delete, or merge existing contacts through this toolkit.
Arcade requests the Google OAuth scopes required by the actions you enable. Searching contacts generally requires contacts.readonly, creating contacts requires contacts, directory searches require directory.readonly, and searching automatically saved contacts may require the corresponding other-contacts read scope. Grant only the scopes your workflow needs.
Yes. Each user authorizes their own Google account through OAuth. Contact searches and contact creation run under that user's authorization, so the agent can access only the account, directory, and permissions that user approved.
Yes. Contextual Access policies can require human approval before consequential actions such as creating a contact. The call proceeds only once the policy's conditions have been satisfied; policies can also redact fields from responses or hide an action from an agent.
Arcade proxies Google Contacts API calls and does not persist your contact content as part of normal tool execution. OAuth credentials are brokered and stored encrypted at rest, and the model never receives raw credentials.
Any framework or client that can connect to Arcade's MCP gateway or tool APIs works, including LangChain, OpenAI Agents, CrewAI, Claude Desktop, Claude Code, Cursor, Visual Studio Code, Microsoft Copilot Studio, and GitHub Copilot.
Arcade refreshes access while the user's Google authorization remains valid. If access is revoked, rotated, or can no longer be refreshed, contact actions stop until the user reconnects their Google account.
Google People API quotas, Workspace directory limits, and account-level limits apply to requests made through Arcade. No separate toolkit-specific rate limit is stated here, so high-volume searches should still respect Google's quotas and handle quota errors or retries.