MCPHOST-SERVER
About
MCP server for ollama, openai, and anthropic bridging
Details
- Author
- JinsonWu
- GitHub stars
- 2
- Downloads
- 127
- Categories
- AI
Jump to
- HTTP-based server mode for interactive LLM requests
- Environment-based configuration via .env file
- Support for mcp.json configuration file
- Command-line flags for configuration
- Default listen port 8115
Clone the repository, copy .example.env to .env and configure it, then build with go build and run with ./mcphost-server. The server starts on the configured port (default 8115) in server mode by default. To use the command-line interface, set server_mode = false in main.go and rebuild.
MCPHOST-SERVER
A server bridging solution for MCPHOST that enables easy communication with local LLM (Ollama) through HTTP requests. This project is originated from mark3labs/mcphost and adds server capabilities for enhanced functionality.
Added Features
- Server mode for HTTP-based communication with local LLM for interactive requests
- Environment-based configuration
Getting Started
1. Clone the repository
2. Copy the example environment file and configure it, see .env for configuration in details:
cp .example.env .env
3. Build the project:
go build
4. Run the server:
./mcphost-server
Usage
Server Mode
The server mode is enabled by default and provides HTTP endpoints for interacting with the local LLM. The server will start on the configured port (default: 8115).
Command Line Mode
To use the command-line interface, set server_mode = false in the main.go file and rebuild the project.
Configuration
The application can be configured through:
- Environment variables (see .example.env)
- mcp.json configuration file
- Command-line flags
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
