Mindtickle MCP
Connect AI clients to Mindtickle with the Model Context Protocol, choose a connection path, and understand access requirements.
The Mindtickle MCP server is Mindtickle’s implementation of the Model Context Protocol (MCP), an open standard that lets AI clients connect to external tools, data sources, and services. It exposes Mindtickle capabilities as tools that MCP-compatible AI clients can call during user conversations, without a separate point-to-point integration for each compatible client.
With the Mindtickle MCP server, your sales reps can ask questions in their AI client (such as “Find the latest battle card for competitor ABC” or “Summarize my last call with Acme”) and receive answers drawn from their Mindtickle content, without leaving the tool they are already working in.
Who this is for
Section titled “Who this is for”Choose a guide based on your role.
| Role | What they do | Articles |
|---|---|---|
| IT admin | Manages which third-party apps and connectors your organization can use inside AI clients such as Claude or ChatGPT. This role requires admin access in the AI client, not the Mindtickle admin site. | Set up the connector |
| Developer | Builds or integrates a custom AI client or agent directly against the Mindtickle MCP server’s API. | Supported tools, then Build a custom client |
| Sales rep | Uses Mindtickle content and tools day-to-day through an AI client. | Connect your account to get set up, then Use cases for what’s possible once connected |
What you can do
Section titled “What you can do”Connect an MCP-compatible AI client to search Mindtickle content, ask Seller Copilot questions, look up CRM records, and create or manage Digital Sales Rooms. Each tool requires a scope, and results respect the user’s existing Mindtickle permissions. See Supported tools for the available tools, their requirements, and their effects.
Connection options
Section titled “Connection options”Both connection options use the same Mindtickle MCP server. Available tools depend on the granted scopes, and results respect the user’s Mindtickle permissions. Choose the option that matches how your AI client authenticates.
| Setup detail | Programmatic integration | Custom connector in an MCP client app |
|---|---|---|
| Who it’s for | Developers and integrators building or configuring an AI client, agent, or automation. | Organization admins and end users of a managed, MCP-compatible AI client application (for example, Claude for Teams or Claude Enterprise). |
| How setup works | Manual configuration in code or a config file (for example, editing Cursor’s mcp.json). |
Point-and-click setup through the client application’s admin UI. Each rep then authorizes their own access through an OAuth consent screen. |
| Authentication flow | OAuth 2.0 client credentials grant. The integrator obtains a client_id and client_secret from Mindtickle, exchanges them for a short-lived JWT, and sets the token in the Authorization header. |
OAuth 2.0 authorization code with consent. Users are redirected to a Mindtickle consent screen, click to authorize, and the client application handles the token internally. |
| Token management | The integrator caches the token, checks expires_in, and obtains a replacement before it expires. |
The AI client manages tokens. Reconnection depends on access revocation, refresh-token expiry, and whether the client registered for a refresh token. |
| Setup guide | Build a custom client | Set up the connector (admin), then Connect your account (rep) |
Authentication and access control
Section titled “Authentication and access control”All requests to the Mindtickle MCP server are authenticated using OAuth 2.0 and short-lived Bearer tokens (JWTs). On every request, the server validates the token and checks the scopes it carries before invoking any tool. This ensures that:
- Only authorized clients can connect.
- Each tool call runs in the context of an identified Mindtickle user.
- Results respect the same permissions and content access policies that apply to that user inside Mindtickle.
- Only the tools matching the granted scopes are exposed and callable.
For detailed information on tools, see Supported tools. For scopes, grant types, the token endpoint, and error codes specific to your connection path, see Build a custom client or Set up the connector.
Availability and access
Section titled “Availability and access”- Credentials and enablement: The integration is not self-service. Mindtickle must enable it on your instance, and any
client_idandclient_secretrequired for the programmatic integration path or for admin-controlled connector setup are issued only by the Mindtickle team. - How to get started: Contact your Customer Success Manager or account team to enable the integration and request credentials.
In this section
Section titled “In this section”- Set up the connector: Configure the connector and scope controls for your organization.
- Connect your account: Authorize your account, use the connector, and troubleshoot access.
- AI clients: Find setup guidance for your AI client.
- Build a custom client: Obtain credentials, authenticate, and call tools from your integration.
- Supported tools: Review tool parameters, response fields, limits, and side effects.
- Use cases: Explore prompts and workflows for content discovery and Digital Sales Rooms.
- Best practices: Apply security, reliability, and integration guidance.