# Mikensey MCP Server

McKinsey-style strategy intelligence for real estate — powered by 53 episodes of Mike DelPrete's [Context podcast](https://library.mikedp.com/).

An open-source [MCP server](https://modelcontextprotocol.io/) that gives Claude (or any MCP-compatible AI) access to real estate industry benchmarks, consulting frameworks, and full-text transcript search. Ask it to analyze your business situation using 2×2 matrices, SCR memos, issue trees, or Porter's Five Forces — grounded in real data from industry leaders.

## Tools

| Tool | What it does |
|------|-------------|
| `mikensey_search` | Full-text search across all 53 transcripts |
| `mikensey_list_episodes` | Browse and filter episodes by guest or topic |
| `mikensey_get_episode` | Read full or partial transcripts |
| `mikensey_get_benchmarks` | Query 26 industry benchmarks (attach rates, costs, growth metrics) |
| `mikensey_get_frameworks` | Browse 10 industry frameworks + 6 consulting frameworks |
| `mikensey_analyze` | Apply a consulting framework (2×2, SCR, Issue Tree, Porter's, Value Chain, MECE) to your situation |
| `mikensey_get_advice` | SCR-structured strategy brief combining search, benchmarks, and frameworks |

## Quick Start

```bash
git clone https://github.com/YOUR_USERNAME/mikensey-mcp-server.git
cd mikensey-mcp-server
npm install
npm run build
```

Add to Claude Desktop (`~/Library/Application Support/Claude/claude_desktop_config.json` on Mac):

```json
{
  "mcpServers": {
    "mikensey": {
      "command": "node",
      "args": ["/path/to/mikensey-mcp-server/dist/index.js"],
      "env": {
        "MIKENSEY_TRANSCRIPT_DIR": "/path/to/your/transcript-files"
      }
    }
  }
}
```

Restart Claude Desktop. You'll see "mikensey" in the tools menu.

## Usage

Just talk to Claude naturally:

- *"Analyze my situation using a 2×2 matrix — I'm a broker losing agents to cloud brokerages"*
- *"Give me an SCR strategy memo — I'm a mortgage LO worried about AI replacing my job"*
- *"Build an issue tree for how a proptech startup can achieve profitability"*
- *"What are the benchmark attach rates for mortgage in real estate?"*
- *"Which episodes cover Zillow's strategy?"*

## What's Inside

**26 industry benchmarks** — specific numbers cited by leaders like Tamir Poleg (Real), Glenn Sanford (eXp), Garth Graham (Stratmore), Justin Messer (Prosperity), and more. Attach rates, origination costs, agent productivity, growth metrics.

**10 industry frameworks** — mental models from the podcast: the 3-Lever Consumer Value Model, the Brokerage Pretzel, the Agent Value Equation, Hand-to-Hand Combat, W-H-Y sales framework, and more.

**6 consulting frameworks** — McKinsey-style analytical tools: 2×2 Matrix, Situation-Complication-Resolution (SCR), Issue Trees (MECE), Porter's Five Forces, Value Chain Analysis, and MECE Breakdowns. Each comes with real estate examples and pre-built templates.

**4 pre-built 2×2 matrices** — Brokerage Business Models, Mortgage Origination Strategy, Proptech Startup Survival, Agent Recruitment & Retention — with companies plotted and insights synthesized from transcript evidence.

**3 pre-built issue trees** — Brokerage Profitability, Mortgage Growth Strategy, Proptech Product-Market Fit — with MECE branches and supporting data.

## Transcript Format

The server expects `.txt` files named like:

```
YYYY-MM-DD_guest-name-episode-title.txt
```

For example: `2024-03-15_varun-krishna-rockets-next-chapter.txt`

## Configuration

| Environment Variable | Default | Description |
|---------------------|---------|-------------|
| `MIKENSEY_TRANSCRIPT_DIR` | `~/Documents/CodingJourney/Mikensey/context-podcast-transcripts` | Path to folder with `.txt` transcript files |

## Development

```bash
npm run dev    # Watch mode with tsx
npm run build  # Compile TypeScript
npm start      # Run compiled server
```

## BYO Transcripts

This server works with any set of transcripts in the expected format. You could adapt it to your own podcast, interview series, or research corpus — just point `MIKENSEY_TRANSCRIPT_DIR` at your files.

## Cost

Zero. Runs entirely locally. No API keys, no external services, no ongoing costs.

## License

MIT