> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getwhys.io/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP integration

> GetWhys plugs into Claude, Cursor, Typeface, and the agents your team has built. Buyer intelligence becomes the layer behind every AI workflow.

Your team already lives in Claude, Cursor, Typeface, Copilot, and the AI agents they've built themselves. The GetWhys MCP server makes GetWhys show up wherever they already work, so buyer intelligence isn't a tab someone opens when they remember to — it's the layer every tool in the stack draws from automatically, every time.

This is the validation layer extended across your entire AI stack.

## What MCP unlocks

Through MCP, any AI tool your team uses can call GetWhys directly. Personas, messaging frameworks, brand voice, and the buyer signal database become callable resources inside Claude, Cursor, Typeface, and any custom agent. A few patterns:

* **AI content generation with built-in buyer context.** Your team's content tool calls into GetWhys for the persona and messaging framework before it writes a draft. The output starts grounded instead of starting from scratch.
* **Agentic research that pulls real interviews.** When an agent is researching a competitive teardown or a market scan, it queries GetWhys for primary buyer research instead of relying on general model knowledge.
* **Persona-aware content scoring inside other tools.** Tools your team has built can call GetWhys to validate copy mid-draft, not just at the end.
* **Always-on buyer intelligence in Slack, your CRM assistant, your internal agents.** One source of truth, accessible anywhere your team works.

## Tools exposed through MCP

| Tool                           | What it does                                                                                                 |
| ------------------------------ | ------------------------------------------------------------------------------------------------------------ |
| `list_personas`                | Returns your published Personas (handle and name). Useful for picking the right one before fetching details. |
| `get_persona`                  | Returns a single Persona's full content: challenges, motivations, KPIs, job titles, watering holes.          |
| `get_all_personas`             | Returns every published Persona in one call.                                                                 |
| `list_messaging_frameworks`    | Returns your published Messaging Frameworks.                                                                 |
| `get_messaging_framework`      | Returns a single Framework's full content.                                                                   |
| `get_all_messaging_frameworks` | Returns every published Framework in one call.                                                               |
| `get_brand_voice`              | Returns your Brand Voice characteristics with do's and don'ts.                                               |
| `query_market_research`        | Asks a buyer research question and returns a synthesized, evidence-grounded answer with credibility signals. |
| `ask_market_research`          | Same as `query_market_research` but also returns structured citations alongside the answer.                  |
| `whoami`                       | Returns the authenticated user and organization context.                                                     |

## Generating an API token

<Steps>
  <Step title="Click Integrations in the left navigation">
    Click **Integrations** in the left navigation
  </Step>

  <Step title="Find the MCP API Tokens card">
    Find the **MCP API Tokens** card
  </Step>

  <Step title="Click Create token and give it a label (e.g., &#x22;Claude Desktop&#x22;, &#x22;Cursor&#x22;, &#x22;Custom Agent&#x22;)">
    Click **Create token** and give it a label (e.g., "Claude Desktop", "Cursor", "Custom Agent")
  </Step>

  <Step title="Copy the token immediately — it's only shown once">
    Copy the token immediately — it's only shown once

    If you lose a token, revoke it and create a new one.
  </Step>
</Steps>

## Configuring your MCP client

Each MCP client configures servers slightly differently. The connection details:

* **Server URL** — provided on the integration card when you create the token
* **Auth header** — use a bearer token in the `Authorization` header:

```text theme={null}
Authorization: Bearer <your-token>
```

Refer to your client's documentation for where to paste these. Claude Desktop, Claude Code, Cursor, and most newer agentic tools support MCP server configuration.

## For PMM and marketing leaders

MCP is how you make your judgment scale. Your defined personas, your messaging framework, and your brand voice become the constraints that every AI workflow in the company runs against — automatically. Sellers don't need to know that the proposal they built in Claude was improved by GetWhys. They just get a better proposal. Your PMMs don't need to change how they work. Buyer intelligence sits behind every tool they already use.

You set the standard once. GetWhys enforces it everywhere your team builds.

## Revoking tokens

Return to the **Integrations** page, find the token in the MCP API Tokens list, and click **Revoke**. The token stops working immediately. Other tokens are unaffected.

## Security

<Warning>
  Tokens are scoped to your user and organization. Treat them like passwords:

  * Don't commit them to source control
  * Don't share them across teammates — generate one per person or per agent
  * Rotate them periodically and after any device change

  If you suspect a token has leaked, revoke it immediately and generate a new one.
</Warning>