dolphin-mcp

NPM Version License: MIT Bun

A powerful Model Context Protocol (MCP) server that provides Dolphin AI semantic code search capabilities to AI applications like Continue.dev, Claude Desktop, and other MCP-compatible clients.

🚀 Features

Semantic Code Search: Find code using natural language queries powered by AI embeddings
Multi-Repository Support: Search across multiple code repositories
Smart Context Retrieval: Get detailed code chunks and file slices
VS Code Integration: Generate editor-friendly file URIs
Robust Error Handling: Graceful fallbacks and informative error messages
Environment Configuration: Flexible API endpoint configuration

🛠️ Installation

Prerequisites

Bun (>= 1.0.0): Install Bun
Dolphin API Server: Running on your specified endpoint

Install the MCP Server

# Install globally bun install -g @plastic-beach/dolphin-mcp # Or install locally in your project bun install @plastic-beach/dolphin-mcp

⚙️ Configuration

Environment Variables

Variable	Required	Default	Description
`DOLPHIN_API_URL`	No	`http://127.0.0.1:7777`	Dolphin API endpoint
`KB_REST_BASE_URL`	No	-	Alternative for DOLPHIN_API_URL
`LOG_LEVEL`	No	`info`	Logging level (debug, info, warn, error)
`SERVER_NAME`	No	`dolphin-mcp`	Server identifier
`SERVER_VERSION`	No	`0.1.0`	Server version

Quick Setup

# Test the server bunx @plastic-beach/dolphin-mcp

📱 AI Application Integration

Continue.dev Configuration

Add this to your config.yaml:

mcpServers: - name: Dolphin-KB command: dolphin-mcp env: DOLPHIN_API_URL: "http://127.0.0.1:7777" connectionTimeout: 30000

Claude Desktop Configuration

Add this to your claude_desktop_config.json:

{ "mcpServers": { "dolphin-kb": { "command": "dolphin-mcp", "env": { "DOLPHIN_API_URL": "http://127.0.0.1:7777" } } } }

Direct Usage

# Start the MCP server directly dolphin-mcp # With custom configuration DOLPHIN_API_URL="https://your-dolphin-instance.com" dolphin-mcp

🧰 Available Tools

The Dolphin MCP server provides these tools to AI applications:

1. search_knowledge

Search codebase semantically using AI embeddings.

Parameters:

query (string): Search query - Required
repos (string[]): Optional repository filters
top_k (number): Number of results (default: 8)

2. fetch_chunk

Get detailed chunk content by ID.

Parameters:

chunk_id (string): Chunk identifier - Required

3. fetch_lines

Get specific file lines by range.

Parameters:

repo (string): Repository name - Required
path (string): File path - Required
start (number): Start line (1-indexed) - Required
end (number): End line (inclusive) - Required

4. get_vector_store_info

Get knowledge base statistics and repository info.

Parameters: None

5. open_in_editor

Generate VS Code URI for opening files.

Parameters:

repo (string): Repository name - Required
path (string): File path - Required
start_line (number): Start line (default: 1)

🔧 Development

Prerequisites

Bun (>= 1.0.0)
TypeScript
Node.js 18+

Setup

# Clone and setup git clone https://github.com/plasticbeachllc/dolphin-mcp.git cd dolphin-mcp # Install dependencies bun install # Build the project bun run build # Run tests bun test # Development mode bun run dev

Project Structure

├── src/ │ ├── cli.ts # CLI entry point │ ├── index.ts # Development entry │ ├── mcp/ │ │ ├── server.ts # MCP server implementation │ │ └── tools/ # MCP tool definitions │ ├── rest/ │ │ └── client.ts # REST API client │ └── util/ # Utilities ├── dist/ # Built files ├── tests/ # Test suites └── README.md

📋 Requirements

Bun: >= 1.0.0
Node.js: 18+
Dolphin API Server: Accessible via HTTP/HTTPS

🐛 Troubleshooting

Common Issues

Server fails to start:

# Check if Dolphin API is accessible curl http://127.0.0.1:7777/health # Verify environment variables echo $DOLPHIN_API_URL # Test with explicit URL DOLPHIN_API_URL="http://127.0.0.1:7777" dolphin-mcp

Connection timeout:

Verify DOLPHIN_API_URL is correct and accessible
Check firewall/network settings
Increase connectionTimeout in your MCP configuration

No search results:

Ensure repositories are indexed in Dolphin
Check repository names match exactly
Verify API credentials and permissions

Logging

The server logs to mcp-bridge/logs/mcp.log with automatic rotation (5MB, 3 files).

🤝 Contributing

Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit changes (git commit -m 'Add amazing feature')
Push to branch (git push origin feature/amazing-feature)
Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔗 Links

NPM Package: https://www.npmjs.com/package/@plastic-beach/dolphin-mcp
GitHub Repository: https://github.com/plasticbeachllc/dolphin-mcp
MCP Specification: https://modelcontextprotocol.io
Continue.dev: https://continue.dev

⚡ Quick Start Commands

# Install bun install -g @plastic-beach/dolphin-mcp # Configure export DOLPHIN_API_URL="http://127.0.0.1:7777" # Start server dolphin-mcp # Test in terminal echo '{"jsonrpc": "2.0", "id": 1, "method": "tools/list"}' | dolphin-mcp

This server cannot be installed

-

security - not tested

A

license - permissive license

-

quality - not tested

How are these scores calculated?

remote-capable server

The server can be hosted and run remotely because it primarily relies on remote services or has no dependency on the local environment.

Provides AI-powered semantic code search across multiple repositories, allowing natural language queries to find code chunks and retrieve specific file contents through Dolphin AI embeddings.

Dolphin MCP

dolphin-mcp

🚀 Features

🛠️ Installation

Prerequisites

Install the MCP Server

⚙️ Configuration

Environment Variables

Quick Setup

📱 AI Application Integration

Continue.dev Configuration

Claude Desktop Configuration

Direct Usage

🧰 Available Tools

1. search_knowledge

2. fetch_chunk

3. fetch_lines

4. get_vector_store_info

5. open_in_editor

🔧 Development

Prerequisites

Setup

Project Structure

📋 Requirements

🐛 Troubleshooting

Common Issues

Logging

🤝 Contributing

📝 License

🔗 Links

⚡ Quick Start Commands

New MCP Servers

MCP directory API