TDQS logo

TDQS

Score how well your MCP tools speak to agents.

freemium Browser Terminal HTTP Programming Research Lab

TDQS is a programming research lab tool built by Glama. It's best for Developers building AI agents and Teams managing Multi-Capability Platforms (MCP). Pricing is freemium.

Pricing

freemium

Audience

Developers building AI agents

Platforms

Community

0%

About TDQS

TDQS is an open framework for scoring how well an MCP (Multi-Capability Platform) tool definition communicates to an AI agent, providing a specification, reference implementation, CLI, and hosted API.

TDQS grades every tool definition on six weighted dimensions, explaining each point, and converts the result into a tier that can be used to gate releases. The scoring is consistent across terminals, CI environments, and Glama's registry, ensuring reliability and comparability.

The scoring process is designed to be deterministic where possible and judged by a model where necessary. It involves four stages: context signals (reading schema and annotations), hard gates (short-circuiting degenerate definitions), a rubric (a model grading six dimensions against published anchors), and post-processing (applying overrides, flags, and smells to generate a final score and tier).

TDQS measures ten dimensions in total: six weighted dimensions for each tool (Purpose Clarity, Usage Guidelines, Behavioral Transparency, Parameter Semantics, Conciseness & Structure, Contextual Completeness) and four equally weighted dimensions for the server as a whole (Disambiguation, Naming Consistency, Tool Count Appropriateness, Completeness). Each score comes with a written justification, and every dimension has a dedicated page with its anchors. The framework offers a playground, a command-line interface (CLI), and a hosted API for scoring, ensuring the same result regardless of the method used.

Key Features

Open framework for tool definition quality scoring
Six weighted dimensions for tool grading
Four equally weighted dimensions for server grading
Deterministic scoring for schema and annotations
AI model-based grading for description quality
Justifications for every score point
Tiered scoring system (A, B, C, D, F)
CLI for local and CI integration
Hosted API for programmatic scoring
Browser-based playground for interactive scoring
Shareable reports with SVG badges
Specification for scoring pipeline and prompts
Reference implementation available on npm and PyPI
Integration with Glama's registry

Pricing

freemium

The hosted API offers thirty calls a day per account for free. No other pricing details are explicitly mentioned on the website.

Who is it for?

Best for

  • Ensuring high-quality tool definitions for AI agents
  • Automating quality checks for API descriptions in CI/CD pipelines
  • Standardizing tool description quality across an organization
  • Improving AI agent tool selection accuracy
  • Providing clear feedback for improving tool documentation

Not ideal for

  • General-purpose API documentation generation
  • AI agent development itself (focuses on tool quality, not agent logic)
  • Non-MCP tool definition scoring

Integrations

npm PyPI GitHub Glama's registry

Community Discussion

Sign in to contribute

No discussions yet. Be the first to share your experience!

Frequently asked questions