anytra by kyoabhiem - MCP Server

Anytra

A Rust-based MCP (Model Context Protocol) server that enhances your prompts using AI to make them clearer and more effective.

What it does

This tool helps you improve your prompts before sending them to AI models. It can:

Make prompts clearer and more specific
Add context and structure
Adjust tone and style
Support different languages
Provide quality assurance
Sequential thinking for step-by-step reasoning

Quick Start

Requirements

Rust 1.75 or newer
OpenRouter API key

Installation & Setup

Clone and build:

git clone <repository-url>
cd anytra
cargo build --release

Get an OpenRouter API key:
- Sign up at openrouter.ai
- Get your API key from the dashboard

Set up environment:

export OPENROUTER_API_KEY=your_api_key_here
# Optional: set model preference
export OPENROUTER_MODEL=openrouter/auto
# Optional: enable sequential thinking by default
export ENABLE_SEQUENTIAL_THINKING=true

Run the server:
```
cargo run -- --log-level info
```

Usage

Basic Enhancement

echo '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"enhance_prompt","arguments":{"prompt":"write code for fibonacci"}}}' | cargo run --quiet --

Sequential Thinking Enhancement

For complex problems that benefit from step-by-step reasoning:

echo '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"enhance_prompt","arguments":{"prompt":"solve this coding problem","goal":"provide step-by-step solution","enable_sequential_thinking":true,"thought_count":3}}}' | cargo run --quiet --

Controlling Sequential Thinking via Environment Variable

You can enable sequential thinking by default for all requests:

export ENABLE_SEQUENTIAL_THINKING=true
# Now all prompts will use sequential thinking unless explicitly disabled
echo '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"enhance_prompt","arguments":{"prompt":"complex problem"}}}' | cargo run --quiet --

Integration

With Windsurf

Update your MCP config file:

{
  "mcpServers": {
    "prompt-enhancer": {
      "command": "/path/to/mcp-prompt-server/target/release/mcp-prompt-server",
      "args": ["--log-level", "info"],
      "env": {
        "OPENROUTER_API_KEY": "your_api_key_here",
        "ENABLE_SEQUENTIAL_THINKING": "true"
      }
    }
  }
}

Features

AI-Powered Enhancement: Uses OpenRouter API for intelligent prompt improvement
Quality Validation: Ensures enhanced prompts meet quality standards
Fallback Support: Works even when AI services are unavailable
Flexible Options: Customize enhancement with goals, styles, tones, and more
Multi-language Support: Enhance prompts in different languages
Sequential Thinking: Enable step-by-step reasoning for complex problem-solving

Command Line Options

--log-level <level>: Set logging level (error, warn, info, debug, trace)
--shutdown-timeout <secs>: Graceful shutdown timeout in seconds (default: 5)

Environment Variables

OPENROUTER_API_KEY: Required for AI enhancement
OPENROUTER_MODEL: Optional model selection (default: openrouter/auto)
OPENROUTER_REFERER: Optional, recommended for routing
OPENROUTER_TITLE: Optional, recommended for routing
ENABLE_SEQUENTIAL_THINKING: Enable sequential thinking by default (true/false, default: true)

Testing

Run tests:

cargo test

Test the server:

echo '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | cargo run --quiet --

Project Structure

anytra/
├── src/
│   ├── domain/           # Core business logic
│   │   ├── models.rs     # Data structures
│   │   ├── llm.rs        # AI provider interface
│   │   ├── validation.rs # Quality checks
│   │   ├── fewshot.rs    # Example prompts
│   │   └── sequential_thinking.rs # Sequential thinking logic
│   ├── usecases/         # Application logic
│   │   └── enhance_prompt.rs
│   ├── infrastructure/   # External services
│   │   ├── config.rs     # Environment configuration
│   │   ├── logger.rs     # Logging setup
│   │   └── providers/
│   │       └── openrouter.rs
│   └── interface/        # MCP server
│       └── mcp/
│           └── server.rs
├── Cargo.toml
└── README.md