Skip to main content
Early access. Reevo MCP is currently available only to organizations enabled by Reevo. Connections from workspaces that have not been enabled fail at authorization. Contact your Reevo representative or reach out through AskReevo support to request access.
Looking for the AI assistant built into Reevo? That’s Ask Reevo. Use Ask Reevo to chat with your data without leaving Reevo. Use the MCP server when you want your own external agent to reach into Reevo.
Through the MCP server, an assistant can work with your:
  • Accounts
  • Contacts
  • Opportunities and pipelines
  • Meetings
  • Tasks
  • Notes
  • Custom objects
  • Email threads
  • Users
Every call runs as the connected user, respects your existing Reevo permissions, and operates on live production data.

Things You Can Ask Your Assistant to Do

Once connected, you can prompt your assistant for things like:
  • Pipeline reviews - “Which open opportunities have not had engagement in 14 days and have a close date this quarter?”
  • Account and contact lookup - “Find all contacts at Stripe and tell me who owns the account.”
  • Activity summaries - “Summarize the last three meetings on the Acme deal.”
  • Pipeline movement - “Move the Globex opportunity to Negotiation and set the amount to $48,000.”
  • Notes and tasks - “Log a note on the Acme opportunity summarizing today’s call, and create a follow-up task for me on Friday.”
  • Custom objects - “List every record in our partner_program custom object that is in the Tier 1 segment.”
The assistant works on live data. Every write action goes straight to your production workspace, and Reevo cannot undo it. Some actions have external effects, for example enrolling contacts in a sequence sends real outreach emails. Review what your assistant proposes before approving any write action.

Setting Up the Connection

Before You Start

You will need an active Reevo account in the workspace whose data you want the assistant to access, plus an MCP-compatible AI client. Reevo MCP works with any client that supports MCP’s streamable-http transport with OAuth 2.0 authorization. This includes Claude Desktop (Custom Connectors), Claude.ai (Connectors), Cursor, and ChatGPT Developer Mode connectors. The MCP server URL is: https://mcp.reevo.ai/mcp

Connect from Claude

  1. Open the Claude desktop app or the web app and sign in.
  2. In the left sidebar, click Customize.
  3. Open the Connectors section.
  4. Click the + button, then click Add custom connector.
  5. Fill in the connector form:
    • Name: Reevo
    • Remote MCP server URL: https://mcp.reevo.ai/mcp
  6. Click Add. Claude opens your browser to Reevo’s login.
  7. Sign in to Reevo with the account whose data you want the assistant to access, then click Authorize to grant access. The authorization page lists each scope the client is requesting. See Permissions and Scopes below for what each scope grants.

Connect from ChatGPT (Developer Mode)

  1. In ChatGPT, open Settings > Apps > Advanced settings and enable Developer Mode.
  2. From the Apps screen, click Create app.
  3. Enter Reevo as the name and https://mcp.reevo.ai/mcp as the server URL.
  4. Save the app, then complete the Reevo sign-in and authorization prompt. The authorization page lists each scope the client is requesting. See Permissions and Scopes below for what each scope grants.

Connect from Other MCP Clients

Any client that supports MCP’s streamable-http transport with OAuth 2.0 can connect, including Cursor. The steps vary by client, but the pattern is the same:
  1. Find the client’s MCP or connector settings and add a new remote server.
  2. Enter https://mcp.reevo.ai/mcp as the server URL.
  3. Complete the Reevo sign-in and click Authorize when your browser opens Reevo’s login page.
Refer to your client’s documentation for where remote MCP servers are configured.

Verify Your Connection

After connecting, confirm everything works with a quick test. Ask your assistant: “Which Reevo user am I connected as?” The assistant should reply with your Reevo user and organization. If it cannot answer, see the Troubleshooting / FAQs section below.

Permissions and Scopes

Reevo MCP uses OAuth 2.0 with granular scopes. When you connect, Reevo’s authorization page shows exactly which scopes the client is requesting, each tied to an object type and an action (read, create, or update). The assistant can only call a tool if you granted the matching scope at sign-in. Scopes follow a verb:resource format. The scopes used by the published tools are:
  • read:account, create:account, update:account
  • read:contact, create:contact, update:contact
  • read:opportunity, create:opportunity, update:opportunity
  • read:pipeline
  • read:meeting
  • read:task, create:task, update:task
  • read:note, create:note
  • read:custom_object, create:custom_object, update:custom_object
  • read:sequence, update:sequence
  • read:email_account
A few read tools reuse the read:account scope rather than a dedicated scope: search_users, search_email_threads, and get_object_schema all require read:account. search_activity_timeline requires the read scope for each activity type it returns (read:meeting, read:task, read:note, read:email_account), all of which are already listed above. whoami requires no scope and is available to any connected user. Your Reevo permissions still apply on top of OAuth scopes. Granting a scope does not let the assistant see or change anything your own account cannot.

The Tool Catalog

The default Reevo MCP catalog publishes 30 tools, each gated by a granular OAuth scope. The assistant can only call a tool if you granted the matching scope at sign-in. Operators can publish a narrower or broader catalog per organization. This catalog is an early release and will continue to change as coverage expands.

Accounts

Contacts

Opportunities and Pipelines

Meetings

Activity

Tasks

Notes

Email

Sequences

Custom Objects and Schema

Workspace


Troubleshooting / FAQs

AskReevo is the conversational assistant built into the Reevo product. The MCP server is what connects an external assistant, such as Claude or ChatGPT, to your Reevo workspace. Use AskReevo to chat with your data inside Reevo, and use the MCP server when you want your own outside agent to read from and write to Reevo.
No. Two layers apply. First, the assistant can only call a tool if you granted the matching OAuth scope when you connected. Second, every call runs as your user and respects your existing Reevo permissions, so the assistant cannot see or change anything your own account cannot.
Yes. At sign-in, Reevo’s authorization page lists each requested scope by object type and action (read, create, update). You grant only the scopes you are comfortable with, and the catalog your organization publishes can be narrower than the default set.
Reevo MCP is in early access and is enabled per organization. If sign-in completes but authorization fails, or the Reevo option never appears, your organization may not have access yet. Contact your Reevo representative or reach out through AskReevo support to request access. If your organization does have access, remove the connector from your AI client and add it again to restart the OAuth flow.
Two layers can block a call. If you did not grant the matching OAuth scope at sign-in, the tool is unavailable; reconnect and grant the scope to enable it. If the scope was granted but your own Reevo role does not allow the action, the call fails the same way it would if you tried it in Reevo directly; ask your Reevo admin about your role’s permissions.
Scopes are set during the OAuth sign-in. To change them, remove the Reevo connector from your AI client, add it again, and grant the new set of scopes on Reevo’s authorization page.
Remove the Reevo connector from your AI client to stop new calls. Your Reevo administrator can also disable MCP access for the organization, which cuts off existing connections on their next call.
Still have questions? Sign in and use Ask Reevo for instant answers or to raise a support ticket.