Skip to main content
Glama
max-paulus

gong-mcp

by max-paulus
README.md
# Gong MCP

MCP integration for Gong API, enabling AI assistants to interact with Gong call recordings, transcripts, and analytics.

## Features

- Authentication via secure Basic Auth
- Access to calls, transcripts, users, and more
- Compatible with Cursor AI assistant
- Includes brief generation capability

## Prerequisites

- Node.js >= 20.0.0
- pnpm (recommended) or npm
- Gong API credentials (Access Key and Secret)

## Setup

1. Clone this repository
2. Install dependencies:
   ```bash
   pnpm install
   ```
3. Build the TypeScript code:
   ```bash
   pnpm build
   ```
4. Configure your global Cursor MCP config at `~/.cursor/mcp.json` with your Gong credentials:
   ```json
   {
     "mcpServers": {
       "gong-mcp": {
         "type": "stdio",
         "command": "/opt/homebrew/bin/node",
         "args": ["/path/to/gong-mcp/build/index.js"],
         "cwd": "/path/to/gong-mcp",
         "environment": {
           "GONG_ACCESS_KEY": "your-access-key",
           "GONG_SECRET": "your-secret"
         }
       }
     }
   }
   ```

## Environment Variables

The following environment variables are required:
- `GONG_ACCESS_KEY`: Your Gong API access key
- `GONG_SECRET`: Your Gong API secret

These can be set either in your environment or in the Cursor MCP config.

## Documentation

- [Main Documentation](docs/gong-mcp.md) - Detailed API and usage documentation
- [Sample Requests](docs/sample_requests.md) - Example API requests and responses

## Development

Available npm scripts:
```bash
pnpm build        # Build the TypeScript code
pnpm dev          # Watch mode for development
pnpm test         # Run all tests
pnpm test:smoke   # Run smoke tests only
pnpm lint         # Run ESLint
```

## Testing

Run the smoke tests to validate your setup:
```bash
pnpm test:smoke
```
Note: Make sure to have `GONG_ACCESS_KEY` and `GONG_SECRET` environment variables set before running tests.

## Usage with Cursor

Once configured in your global Cursor MCP settings, the AI assistant can:
- Retrieve call recordings and transcripts
- Search for calls by date range or participants
- Access user information
- Generate briefs for deals/accounts
- Analyze call patterns and metrics 

TDQS

C2.4/5.0

Scored across 6 tools

Disambiguation4/5

The tools have distinct purposes: generating briefs, retrieving calls, GDPR activities, listing users, filtering calls, and downloading transcripts. There is some potential overlap between getv2callsbyid and postv2callsextensive (both involve call data), but their descriptions clarify that one retrieves a single call while the other filters lists, reducing confusion.

Naming Consistency2/5

The naming is inconsistent, mixing HTTP methods (get, post) with descriptive suffixes in a non-uniform pattern. For example, getv2askanythinggeneratebrief includes an action verb 'generate' in the name, while others like getv2users do not. This lack of a clear, predictable naming convention makes the set harder to navigate.

Tool Count4/5

With 6 tools, the count is reasonable for a Gong CRM integration server. It covers key areas like calls, users, and data privacy without being overwhelming. However, it feels slightly thin for comprehensive CRM operations, such as missing update or delete functionalities.

Completeness3/5

The toolset covers core read operations for calls, users, and GDPR activities, but lacks obvious CRUD completeness. There are no tools for creating, updating, or deleting resources, which are essential for full CRM workflows. This creates gaps that agents might struggle with for comprehensive task handling.

Maintenance

ActivityInactive
ResponsivenessNo issues