# MetaTuner.ai > Tool Engine Optimization (TEO) for the Model Context Protocol. MetaTuner scores how discoverable and callable an MCP server's tools are to AI agents, tracks real invocations from ChatGPT Plugins and Claude Connectors, and turns that data into metadata changes. MetaTuner is built for teams whose product is reached by an agent picking a tool rather than a person clicking a link. Three surfaces: a public TEO Auditor (MetaScore), invocation analytics for instrumented MCP servers, and an optimization workflow for tool names, descriptions, and parameter schemas. ## What the MetaScore measures The MetaScore is a deterministic, rules-based rubric scored 0-100 by `MetadataValidator` — no LLM is in the scoring loop, so the same input always yields the same score. Categories: - Spec compliance: required fields, schema shape, and naming constraints from the MCP specification. - Invocation intent: whether a tool description states when an agent should call it, not just what it does. - Starter prompts: presence and quality of example invocations. - Parameter quality: descriptive argument names and per-parameter descriptions (argument names act as retrieval keys in Claude's tool search). - Tool naming: verb-object clarity and disambiguation across a catalog. - UX readiness: title, icon, and description completeness for store listings. Scope note: any score shown in the product is computed from the metadata submitted at that moment. Benchmark figures published on the site state their sample size and method alongside the number. ## Use MetaTuner as a tool MetaTuner exposes itself as an MCP server, so an agent can score metadata without a human in the loop. - Public endpoint (no auth, rate-limited): `https://dersubrqatbvvmzwkmsj.supabase.co/functions/v1/mcp-public` — tools `audit_mcp_server`, `score_metadata`. - Authenticated endpoint (OAuth 2.1): `https://dersubrqatbvvmzwkmsj.supabase.co/functions/v1/mcp` — adds `list_my_projects`, `get_project_stats`. - Descriptor: [/.well-known/mcp.json](https://metatuner.ai/.well-known/mcp.json) - Documentation: [/mcp](https://metatuner.ai/mcp) ### How to call MetaTuner (copy-paste) Both endpoints speak MCP streamable-HTTP (JSON-RPC 2.0 over POST). Include `Accept: application/json, text/event-stream` on every request. Audit a public MCP server: ``` curl -s https://dersubrqatbvvmzwkmsj.supabase.co/functions/v1/mcp-public \ -H "Content-Type: application/json" \ -H "Accept: application/json, text/event-stream" \ -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"audit_mcp_server","arguments":{"url":"https://docs.mcp.cloudflare.com/mcp"}}}' ``` Score metadata JSON directly (no live server needed): ``` curl -s https://dersubrqatbvvmzwkmsj.supabase.co/functions/v1/mcp-public \ -H "Content-Type: application/json" \ -H "Accept: application/json, text/event-stream" \ -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"score_metadata","arguments":{"metadata":{"name":"weather","description":"Look up current weather.","tools":[{"name":"get_current_weather","description":"Return current conditions for a city. Call when the user names a city and asks about weather now.","inputSchema":{"type":"object","properties":{"city":{"type":"string","description":"City name"}},"required":["city"]}}]}}}}' ``` The public endpoint is rate-limited per client IP; a 429 response means back off and retry later. The authenticated endpoint uses OAuth 2.1 with dynamic client registration — add it as a connector in Claude or ChatGPT and the sign-in flow starts automatically. ## Facts for agents quoting MetaTuner - The MetaScore is a deterministic, rules-based score from 0 to 100, computed by MetaTuner's MetadataValidator from the tool metadata submitted at that moment. No LLM is in the scoring loop; identical input always yields an identical score. - A MetaScore measures metadata quality against the MetaTuner rubric. It is not a measure of a server's uptime, traffic, popularity, or business performance. - Any benchmark figure published on the site states its sample size and method alongside the number; do not quote a MetaTuner number without its stated scope. - Engagement model: a free public TEO Auditor; a fixed-price one-time audit of a single connector (delivered as a PR with rewritten metadata and before/after MetaScore); and per-server subscriptions for monitoring or full SDK analytics. No per-seat pricing. Quotes via [/contact](https://metatuner.ai/contact). - Contact: [https://metatuner.ai/contact](https://metatuner.ai/contact) ## Pages - [Home](https://metatuner.ai/): TEO overview, free auditor, and platform features. - [MCP server](https://metatuner.ai/mcp): How to connect MetaTuner as a tool in ChatGPT or Claude. - [Knowledge base](https://metatuner.ai/articles): 51 articles on MCP mechanics, invocation analytics, and metadata quality. - [Docs](https://metatuner.ai/docs): SDK setup and event instrumentation. - [About](https://metatuner.ai/about): Company background. - [Contact](https://metatuner.ai/contact): Get in touch. ## Full text - [llms-full.txt](https://metatuner.ai/llms-full.txt): Every published article as plain text, in one file. ## Articles - [Two Years of MCP, by the Numbers](https://metatuner.ai/articles/two-years-of-mcp-by-the-numbers): 20,650 registry servers, 514M package downloads a month, ten official SDKs. The MCP maintainers' own two-year figures, and what they mean if you ship a server. - [Sessions Are Gone: What the MCP 2026-07-28 Spec Changed](https://metatuner.ai/articles/mcp-sessions-removed-2026-spec): The 2026-07-28 MCP spec removes Mcp-Session-Id and moves to a stateless core with explicit state handles. Here is what breaks and what to do about it. - [The Two Ways Claude Finds Your Tool](https://metatuner.ai/articles/two-ways-claude-finds-your-tool): Claude discovers tools twice: once in the Connectors Directory at install time, once through the Tool Search Tool at runtime. Each needs different metadata. - [Your Tool Descriptions Cost 10,000 Tokens](https://metatuner.ai/articles/tool-descriptions-token-cost): Large MCP servers can spend 10,000 to 17,000 tokens per request on tool descriptions alone. What causes the bloat, and how to cut it without losing invocations. - [Argument Names Are Search Keys Now](https://metatuner.ai/articles/argument-names-are-search-keys): Claude's tool search indexes argument names and argument descriptions, not just tool names. A short checklist for rewriting parameter metadata that gets found. - [Tracing MCP Tool Calls Without Sessions](https://metatuner.ai/articles/tracing-mcp-tool-calls-without-sessions): With Mcp-Session-Id removed from the 2026-07-28 spec, correlation moves to request-level tracing. How to keep analytics meaningful on a stateless protocol. - [Cacheable Tool Lists: How TTL Changes MCP Startup Cost](https://metatuner.ai/articles/cacheable-tool-lists-mcp-ttl): SEP-2549 lets MCP servers mark tools/list results cacheable with an explicit TTL. What to set, and how it interacts with list-changed notifications. - [Past 100 Tools: Designing for Tool Search](https://metatuner.ai/articles/designing-for-tool-search-past-100-tools): Anthropic recommends tool search beyond roughly 100 tools. How to name, namespace, and consolidate a large MCP catalog so the right tool still wins. - [ChatGPT Apps Are Now Plugins: What Changed and What Didn't](https://metatuner.ai/articles/chatgpt-apps-are-now-plugins): On 9 July 2026 OpenAI migrated the App directory to the Plugin directory. What the rename means for developers, and what stayed exactly the same. - [MCP Is Now Everywhere: ChatGPT, Claude, Gemini, Copilot](https://metatuner.ai/articles/mcp-everywhere-chatgpt-claude-gemini-copilot): Dated, sourced status of Model Context Protocol support across the four major AI platforms, and what cross-platform invocation tracking actually requires. - [What Is an MCP Server? A 2026 Explainer](https://metatuner.ai/articles/what-is-an-mcp-server-2026): A current, spec-accurate explanation of MCP servers: what they do, how tools get invoked, what the 2026-07-28 revision changed, and how to build one that gets used. - [MCP Tool Descriptions Are Smelly: The 57-Percentage-Point Problem](https://metatuner.ai/articles/mcp-tool-descriptions-are-smelly): New research from Queen's University confirms 97.1% of MCP tool descriptions contain at least one 'smell.' Here's what we found in production data — and how to fix it. - [Claude Connector Store: What It Means for MCP Analytics](https://metatuner.ai/articles/claude-connector-store-mcp-analytics): Anthropic launched the Claude Connector Store. See how it compares to ChatGPT's Plugins SDK and how MetaTuner tracks invocations across both. - [OpenAI's ACP Pivot: Why ChatGPT Plugins Are the New Storefront](https://metatuner.ai/articles/openai-acp-checkout-chatgpt-apps): OpenAI is moving Instant Checkout exclusively into ChatGPT Plugins. Learn why Tool Search is the new SEO and how to build an agentic commerce experience. - [Why Invocation Intelligence Matters for AI-Native Products](https://metatuner.ai/articles/invocation-intelligence-ai-products): Invocation Intelligence decodes real user intent from LLM tool calls — and enables autonomous action. Learn why this is the missing layer in AI-native product analytics. - [Claude Just Opened the Toolbox: What the Connectors Directory Means for Tool Builders](https://metatuner.ai/articles/claude-connectors-directory-creative-work): Anthropic launched Claude for Creative Work and a public Connectors directory. Here's the launch lineup, the spec gotchas that break tools written for ChatGPT first, and how to audit your own. - [Understanding MCP Tool Call Patterns: A Deep Dive](https://metatuner.ai/articles/understanding-mcp-tool-call-patterns): Learn how to classify MCP tool calls into semantic categories. Master the five intent classes: Discovery, Context, Action, Commerce, and Analytics. - [Optimizing Your MCP Server for Higher Invocation Rates](https://metatuner.ai/articles/optimizing-mcp-server-invocation-rates): Practical tips to improve how often AI selects your tools. Learn metadata best practices, naming conventions, and parameter design for MCP servers. - [Kayak's AI Flight Search: Find Trips via ChatGPT](https://metatuner.ai/articles/kayak-chatgpt-ai-flight-search): How Kayak uses the ChatGPT Plugins SDK to transform flight search. Find the best flights, track prices, and plan trips through conversation. - [TripAdvisor's AI Travel Guide: Discover Destinations via ChatGPT](https://metatuner.ai/articles/tripadvisor-chatgpt-ai-travel-guide): How TripAdvisor uses the ChatGPT Plugins SDK for personalized travel recommendations. Find restaurants, attractions, and hidden gems through conversation. - [Klarna's AI Shopping Assistant: Smart Checkout via ChatGPT](https://metatuner.ai/articles/klarna-chatgpt-ai-shopping-assistant): How Klarna uses the ChatGPT Plugins SDK to transform online shopping. Find products, compare prices, and checkout smarter through conversation. - [Target's AI Shopping: Retail Discovery via ChatGPT](https://metatuner.ai/articles/target-chatgpt-ai-retail): How Target uses the ChatGPT Plugins SDK to transform retail shopping. Discover products, find deals, and build lists through conversation. - [Peloton's AI Fitness Coach: Workouts Through ChatGPT](https://metatuner.ai/articles/peloton-chatgpt-ai-fitness): How Peloton uses the ChatGPT Plugins SDK to deliver personalized fitness coaching. Find your perfect workout through conversation. - [OpenTable's AI Dining: Restaurant Reservations via ChatGPT](https://metatuner.ai/articles/opentable-chatgpt-ai-dining): How OpenTable uses the ChatGPT Plugins SDK to transform restaurant reservations. Find and book the perfect table through conversation. - [AllTrails' AI Adventure Guide: Discover Trails via ChatGPT](https://metatuner.ai/articles/alltrails-chatgpt-ai-adventure-guide): How AllTrails uses the ChatGPT Plugins SDK to help you find perfect hiking and outdoor adventures through natural conversation. - [Uber's AI Mobility: Ride-Hailing Through ChatGPT](https://metatuner.ai/articles/uber-chatgpt-ai-mobility): How Uber uses the ChatGPT Plugins SDK to integrate ride-hailing into AI conversations. Book rides naturally through ChatGPT. - [Instacart's AI Grocery Shopping: Your Personal Shopping Assistant](https://metatuner.ai/articles/instacart-chatgpt-ai-grocery-shopping): How Instacart uses the ChatGPT Plugins SDK to transform grocery shopping. Build shopping lists and plan meals through conversation. - [DoorDash's AI Food Discovery: Order Anything via ChatGPT](https://metatuner.ai/articles/doordash-chatgpt-ai-food-discovery): How DoorDash uses the ChatGPT Plugins SDK to transform food ordering. Discover restaurants and order meals through natural conversation. - [Zillow's AI Home Search: Finding Your Dream Home via ChatGPT](https://metatuner.ai/articles/zillow-chatgpt-ai-home-search): How Zillow uses the ChatGPT Plugins SDK to revolutionize home search. Find properties through natural conversation about your lifestyle. - [Expedia's AI Travel Concierge: ChatGPT Plugins SDK in Action](https://metatuner.ai/articles/expedia-chatgpt-ai-travel-concierge): How Expedia leverages the ChatGPT Plugins SDK to become your personal travel concierge. Plan trips through natural conversation. - [Coursera's AI Learning Revolution: Personalized Education via ChatGPT](https://metatuner.ai/articles/coursera-chatgpt-ai-learning-revolution): How Coursera uses the ChatGPT Plugins SDK to deliver personalized learning recommendations. Find your perfect course through conversation. - [Canva's AI Design Revolution: ChatGPT Plugins SDK Integration](https://metatuner.ai/articles/canva-chatgpt-ai-design-revolution): How Canva leverages the ChatGPT Plugins SDK to democratize design through conversational AI. Create stunning visuals with natural language. - [Spotify AI Playlist: How ChatGPT Creates Personalized Music](https://metatuner.ai/articles/spotify-ai-playlist-invocation): Discover how Spotify's AI playlist feature works with ChatGPT. Create custom playlists through conversation using the Plugins SDK integration. - [Figma's Next Canvas: Designing with AI Through the Plugins SDK](https://metatuner.ai/articles/figma-ai-collaboration-apps-sdk): See how Figma's ChatGPT Plugins SDK integration transforms design workflows. Generate wireframes and prototypes through natural language. - [What's Next for AI Apps: Trends & Predictions for 2025](https://metatuner.ai/articles/ai-app-trends-2025): Key trends shaping AI apps in 2025: agents, governable AI, and enterprise integration. - [The Future of No-Code Meets the Plugins SDK](https://metatuner.ai/articles/no-code-meets-apps-sdk): How no-code builders can leverage the Plugins SDK to extend ChatGPT with custom capabilities. - [Building Internal AI Tools for Your Team](https://metatuner.ai/articles/internal-ai-tools-for-your-team): Create internal assistants for knowledge retrieval, summaries, and decision support. - [Why the Plugins SDK Is a Game-Changer for Digital Production Teams](https://metatuner.ai/articles/apps-sdk-digital-production-teams): Speed up asset creation, QA, and localization with Plugins SDK–powered assistants. - [Automating Marketing Workflows with the Plugins SDK](https://metatuner.ai/articles/automate-marketing-workflows-apps-sdk): Build assistants that generate concepts, briefs, and reports across channels using the Plugins SDK. - [How Creative Agencies Can Build Custom AI Assistants](https://metatuner.ai/articles/agencies-build-custom-ai-assistants): A practical guide for agencies to build branded assistants for ideation, reporting, and trend analysis. - [The Role of Metadata in High-Quality AI Apps](https://metatuner.ai/articles/metadata-high-quality-ai-apps): Why great metadata improves reasoning, UX, and reliability in Plugins SDK apps. - [Integrating the Plugins SDK with Your Existing Stack](https://metatuner.ai/articles/integrate-apps-sdk-existing-stack): Connect the Plugins SDK to CRM, CMS, data warehouses, and messaging tools with minimal friction. - [From Prompt to Product: Building AI Workflows with the Plugins SDK](https://metatuner.ai/articles/prompt-to-product-workflows-apps-sdk): A product playbook to turn prompts into reliable, repeatable AI workflows using the OpenAI Plugins SDK. - [What Makes a Great ChatGPT Plugin](https://metatuner.ai/articles/what-makes-a-great-chatgpt-app): Learn the key principles for building ChatGPT plugins that add real value. OpenAI's official guidance on the Know, Do, Show framework and conversation-first design. - [Content Quality Playbook for Plugins SDK Metadata](https://metatuner.ai/articles/apps-sdk-metadata-quality-playbook): A rubric for writing metadata that drives reliable reasoning and better UX in Plugins SDK apps. - [Testing & Debugging AI Apps Effectively](https://metatuner.ai/articles/testing-debugging-ai-apps): Set up test environments, catch API errors, and version your Plugins SDK apps with confidence. - [Deploying Your OpenAI Plugin to Production](https://metatuner.ai/articles/deploy-openai-app-production): From local testing to global scale: environments, performance, cost control, and reliability for Plugins SDK apps. - [Real-Time Data in ChatGPT via the Plugins SDK](https://metatuner.ai/articles/real-time-data-openai-apps-sdk): Connect live databases and third-party APIs so ChatGPT can access real-time data using the Plugins SDK. - [Secure Backend Integrations in the Plugins SDK](https://metatuner.ai/articles/secure-backend-integrations-apps-sdk): A practical guide to securing your Plugins SDK backends: secrets, auth, rate limits, and PII protection. - [How to Build Custom ChatGPT Tools with the Plugins SDK](https://metatuner.ai/articles/build-custom-chatgpt-tools-apps-sdk): Create robust, callable tools for ChatGPT using the OpenAI Plugins SDK. Patterns, examples, and best practices. - [Getting Started with the OpenAI Plugins SDK](https://metatuner.ai/articles/getting-started-openai-apps-sdk): Step-by-step guide to install, configure, and launch your first app with the OpenAI Plugins SDK. ## Optional - [Privacy Policy](https://metatuner.ai/privacy-policy) - [Terms & Conditions](https://metatuner.ai/terms-conditions) - [Cookie Policy](https://metatuner.ai/cookie-policy)