Skip to Content
Docs are evolving — expect frequent updates.
@kruncherai/kruncher-mcp

@kruncherai/kruncher-mcp

Model Context Protocol (MCP) server for the Kruncher  Integration API. Enables AI assistants to interact with your Kruncher deal flow, portfolio companies, and analysis data.

Table of Contents

Features

  • Full API Coverage: All 22 Kruncher Integration API endpoints exposed as MCP tools
  • Structured Error Handling: Detailed error responses with codes, retry hints
  • Type-Safe: Built with TypeScript and Zod validation
  • AI-Optimized: Rich tool descriptions to help AI assistants use tools effectively

Requirements

Quick Start

CODE
# One-liner to set up with Claude Desktop
KRUNCHER_API_KEY="your-api-key" npx @kruncherai/kruncher-mcp@latest setup-claude

Then restart Claude Desktop.

Installation & Configuration

Claude Desktop

Option 1: Automatic Setup

CODE
KRUNCHER_API_KEY="your-api-key" npx @kruncherai/kruncher-mcp@latest setup-claude

This automatically configures Claude Desktop. Restart Claude Desktop after running.

Option 2: Manual Configuration

Edit your Claude Desktop config file:

PlatformConfig Path
macOS~/Library/Application Support/Claude/claude_desktop_config.json
Windows%APPDATA%\Claude\claude_desktop_config.json
Linux~/.config/Claude/claude_desktop_config.json

Add the following configuration:

CODE
{
  "mcpServers": {
    "kruncher": {
      "command": "npx",
      "args": ["-y", "@kruncherai/kruncher-mcp@latest"],
      "env": {
        "KRUNCHER_API_KEY": "your-api-key-here"
      }
    }
  }
}

Restart Claude Desktop to apply changes.

Claude Code (CLI)

Add to your Claude Code MCP settings (~/.claude/settings.json):

CODE
{
  "mcpServers": {
    "kruncher": {
      "command": "npx",
      "args": ["-y", "@kruncherai/kruncher-mcp@latest"],
      "env": {
        "KRUNCHER_API_KEY": "your-api-key-here"
      }
    }
  }
}

Or add to a project-specific .claude/settings.json file.

Cursor

Add to your Cursor MCP configuration (~/.cursor/mcp.json):

CODE
{
  "mcpServers": {
    "kruncher": {
      "command": "npx",
      "args": ["-y", "@kruncherai/kruncher-mcp@latest"],
      "env": {
        "KRUNCHER_API_KEY": "your-api-key-here"
      }
    }
  }
}

Restart Cursor to apply changes.

Windsurf

Add to your Windsurf MCP configuration (~/.windsurf/mcp.json):

CODE
{
  "mcpServers": {
    "kruncher": {
      "command": "npx",
      "args": ["-y", "@kruncherai/kruncher-mcp@latest"],
      "env": {
        "KRUNCHER_API_KEY": "your-api-key-here"
      }
    }
  }
}

Restart Windsurf to apply changes.

VS Code + Continue

If you’re using Continue  extension in VS Code, add to your Continue config (~/.continue/config.json):

CODE
{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "npx",
          "args": ["-y", "@kruncherai/kruncher-mcp@latest"],
          "env": {
            "KRUNCHER_API_KEY": "your-api-key-here"
          }
        }
      }
    ]
  }
}

OpenAI (Custom GPTs / Assistants)

OpenAI doesn’t natively support MCP, but you can use an MCP-to-HTTP bridge:

Option 1: Using mcp-proxy

  1. Install and run mcp-proxy:
CODE
npm install -g @anthropics/mcp-proxy
 
# Start the proxy server
KRUNCHER_API_KEY="your-api-key" mcp-proxy --port 3000 -- npx @kruncherai/kruncher-mcp@latest

  1. The proxy exposes the MCP tools as HTTP endpoints at http://localhost:3000

  2. Configure your OpenAI Assistant/GPT to call these endpoints as custom actions

Option 2: Direct API Integration

For OpenAI function calling, you can use the Kruncher REST API directly. See the Kruncher API documentation  for OpenAPI spec.

Other MCP Clients

For any MCP-compatible client, use these connection details:

SettingValue
TransportSTDIO
Commandnpx
Arguments["-y", "@kruncherai/kruncher-mcp@latest"]
EnvironmentKRUNCHER_API_KEY=<your-key>

Available Tools

Projects

ToolDescription
list_projectsList all projects (companies) with pagination
search_projectsSearch projects by keywords
get_portfolio_projectsGet all portfolio companies
create_projectCreate a new project and optionally start analysis
create_project_from_htmlCreate project from HTML content

Analysis

ToolDescription
add_analysisStart a new analysis on an existing project
get_analysis_detailGet detailed analysis information
update_analysis_entityUpdate a specific entity value in an analysis
get_reportGet the full analysis report

Pipeline Management

ToolDescription
add_commentAdd a comment to a project
change_statusChange project pipeline status/phase/stage
set_criteriaSet or update project criteria/scoring
add_voteAdd a vote/rating to a project

Project Lists

ToolDescription
list_project_listsGet all project lists
assign_project_to_listAdd/remove a project from a list

External Mappings

ToolDescription
list_mappingsList all external ID mappings
get_mapping_by_external_idFind mapping by external system ID
get_mapping_by_kruncher_idFind mapping by Kruncher ID
create_mappingCreate an external ID mapping

Integrations

ToolDescription
push_to_crmPush analysis to Attio or Notion
process_affinity_dailySync Affinity list entries
process_affinity_orgProcess single Affinity organization
process_messageProcess inbound email/message

Research & AI

ToolDescription
research_personResearch a person by name and company
chatAsk the Kruncher AI agent questions

Usage Examples

Once configured, you can ask your AI assistant to:

Search and Explore

CODE
"Show me all projects in my pipeline"
"Search for companies in the fintech space"
"Get the full analysis report for Acme Corp"

Create and Manage

CODE
"Create a new project for TechStartup Inc with website techstartup.com"
"Add a comment to the Acme Corp project saying 'Followed up via email'"
"Move the TechStartup project to the 'Due Diligence' stage"

Research and Analysis

CODE
"Research John Smith, CEO of TechStartup Inc"
"What's the revenue growth for companies in my portfolio?"
"Compare the analysis scores of my top 5 deals"

Integrations

CODE
"Push the Acme Corp analysis to Notion"
"Sync my Affinity inbox list"

Environment Variables

VariableRequiredDefaultDescription
KRUNCHER_API_KEYYes-Your Kruncher API key
KRUNCHER_BASE_URLNohttps://api.kruncher.aiAPI base URL (for testing)

Troubleshooting

Server won’t start

  1. Check Node.js version: Requires Node.js 18+

    CODE
    node --version

  2. Verify API key: Ensure KRUNCHER_API_KEY is set correctly

  3. Clear npx cache (if upgrading):

    CODE
    rm -rf ~/.npm/_npx

“import: command not found” error

This indicates an old cached version. Clear the npx cache:

CODE
rm -rf ~/.npm/_npx

Tools not appearing

  1. Restart your MCP client (Claude Desktop, Cursor, etc.)
  2. Check the MCP logs:
    • macOS: ~/Library/Logs/Claude/mcp*.log
    • Windows: %APPDATA%\Claude\logs\mcp*.log

Connection timeouts

Some operations (person research, chat) can take longer. The default timeout is 60 seconds, with 120 seconds for long-running operations.

Debug mode

Run the server directly to see debug output:

CODE
KRUNCHER_API_KEY="your-key" npx @kruncherai/kruncher-mcp@latest

Development

Local Development

CODE
# Clone the repository
git clone https://github.com/kruncherai/kruncher-mcp.git
cd kruncher-mcp
 
# Install dependencies
npm install
 
# Build
npm run build
 
# Run locally
KRUNCHER_API_KEY="your-key" node dist/cli.js

Project Structure

CODE
src/
├── cli.ts              # Entry point with CLI commands
├── stdio.ts            # MCP server setup
├── tools.ts            # MCP tool definitions
├── kruncherClient.ts   # Kruncher API client
└── setupClaude.ts      # Claude Desktop auto-configuration

Building

CODE
npm run build

Publishing

CODE
npm publish --access public

License

MIT License - see LICENSE for details.

Last updated on
ZapierResearch Person