Glama

Glama is a registry, inspector, and gateway platform built for Model Context Protocol servers and connectors. It operates as a superset of the official MCP Registry, indexing tens of thousands of servers, connectors, and tools that are continuously rebuilt and scored for safety and quality. The platform includes a browser-based inspector that lets users test tools in an ephemeral sandbox without local installation. Additionally, it offers a gateway layer that handles authentication, logging, and access control for remote services.

• Browser-based MCP inspector with an ephemeral sandbox to list tools, call them with structured inputs, and handle OAuth flows, bearer tokens, and custom headers
• Comprehensive registry indexing thousands of MCP servers, connectors, and tools with continuous rebuilding and quality-and-safety scoring
• Gateway for remote MCP services that provides authentication, logs, and access control
• One-click deployment and local installation options for open-source MCP servers
• Unified AI Gateway that integrates access to models from OpenAI, Anthropic, Google, DeepSeek, Mistral, and xAI
• Hosted MCP connectors that proxy services like Linear, Stripe, and PostHog into clients like Claude, ChatGPT, Cursor, and VS Code
• Per-project memory, file uploads, web fetch, and web search tools available on higher tiers

Pros

• Provides instant in-browser testing for MCP servers without requiring local installation or manual environment setup
• Maintains a large registry of maintainer-verified, continuously rebuilt servers equipped with quality and safety scores
• Simplifies remote service integration by acting as a single gateway endpoint compatible with multiple AI clients
• Offers a free hosting tier specifically for maintainers of open-source MCP servers

Cons

• Open-source MCP servers are free to browse and install locally, but hosting them on Glama requires a paid plan unless explicitly covered by the open-source hosting waiver
• Paid plans have defined limits on log retention periods, such as 30 days for Starter, 90 days for Pro, and 180 days for Business
• Additional fees apply if a user exceeds the allocated number of fast hosted MCP servers or requires extra log increments on a given plan
• Persistent storage options and advanced features like custom exports and priority support depend heavily on selecting the correct subscription tier

Glama fits developers, teams, and enterprises working with Model Context Protocol servers who need a centralized place to discover, test, and manage MCP connectors and tools. It suits developers building AI agents who want to route requests through a unified API gateway and test server endpoints quickly inside an ephemeral browser sandbox. It is a poor fit for teams not using AI clients or workflows reliant on the Model Context Protocol, or those requiring entirely on-premises self-hosted registry infrastructure without external cloud dependencies.

Glama follows a freemium model with a free tier and three paid subscription plans alongside a free open-source hosting option. The Starter plan costs $9 per month and includes $9 in AI credits, 3 fast hosted MCP servers, 100k logs per month with 30 days retention, and unlimited MCP connectors, with additional servers priced at $4 each. The Pro plan costs $26 per month, providing $26 in AI credits, 10 fast hosted MCP servers, file uploads, shared projects with persistent memory and web fetch or search tools, 100k logs per month with 90 days retention, and additional servers at $3 each. The Business plan costs $80 per month, offering $80 in AI credits, 30 fast hosted MCP servers, custom exports, priority support, 100k logs per month with 180 days retention, and additional servers at $2 each. Hosting and running an MCP server on Glama is free if the server is open-source.

Unlike the official MCP Registry, Glama functions as an expanded superset that actively indexes, scans, and scores every server for maintainer verification, safety, and quality. Instead of only listing endpoints, it integrates an ephemeral browser-based inspector that allows users to test tool calls, inspect raw JSON-RPC traffic, and handle complex authentication flows directly in the browser without local setup. Furthermore, it combines this registry with a managed gateway layer for authentication, access control, and unified AI model access across multiple major providers.

Claude, ChatGPT, Cursor, VS Code, Linear, Stripe, PostHog, OpenAI, Anthropic, Google, DeepSeek, Mistral, and xAI.