Features

Token Efficiency

Token usage minimization while remaining human-readable by TOON

Deterministic Data

Data generation is deterministic and reproducible

MCP protocol

Support for latest protocol version 2025-11-25

Transport

Streamable HTTP in Stateless mode

Validation

Extensive input schema validation

Try it

For stdio or local development see complete instructions in GitHub

System Philosophy

Prompt as an AI Workflow

The /post_engagement_report prompt acts as an AI workflow that guides the user through a structured data analysis process.

User interaction flow:

            1. User types: /post_engagement_report
            2. System prompts: "Enter postId (1-100):"
            3. User enters: "5"
            4. System sends the pre-formatted prompt to the AI
            5. AI executes using available tools

Tools

Prompts define the strategy, while tools provide the execution capabilities. The following tools are used to retrieve and aggregate comment-related metrics:

          get_post_comments
          comments_per_view

Resource Templates

Resources are the data the analysis runs on. URI-style templates are used to reference specific resources by their unique identifiers.

          resource://users/{id}
          resource://posts/{id}

Data Structure

The data model is hierarchical. A user has multiple posts, and each post can contain multiple comments.

            users
            └─ posts
              └─ comments

⚠️ Early development

This project is in an early stage and under active development. Feature requests are welcome!