Smithery vs Glama vs Agensi: MCP Server Directories Compared (2026)
Comparing the three main MCP directories: Smithery, Glama, and Agensi. What each offers, how they differ, and which one fits your needs.
Three platforms have emerged as the main places to find MCP servers: Smithery, Glama, and Agensi. Each takes a different approach, and understanding the differences helps you find what you need faster.
Quick overview
Smithery is a public MCP server registry with the largest catalog. It focuses on breadth — listing as many servers as possible with basic metadata and installation instructions.
Glama is another MCP registry that emphasizes a clean browsing experience and categorization. It's well-organized and easy to navigate.
Agensi combines an MCP directory with a SKILL.md skills marketplace. It's the only platform that covers both MCP servers (tool access) and SKILL.md skills (agent behavior). Agensi also runs automated security scans on all listed items.
Catalog size
Smithery has the largest catalog, indexing thousands of MCP servers from across the ecosystem. Glama has a curated subset. Agensi's MCP directory is smaller but growing, with a focus on quality over quantity.
The raw numbers are less meaningful than they seem. Most MCP servers are experimental or abandoned. A registry with 100 maintained, well-documented servers is more useful than one with 5,000 entries where 90% are broken.
Security and quality
This is where the platforms diverge most:
Agensi runs an automated 8-point security scan on every listing. It checks for excessive permissions, suspicious dependencies, hardcoded credentials, and data exposure risks. Items that fail the scan aren't listed.
Smithery and Glama are registries — they index servers but don't perform security reviews. You're responsible for vetting servers yourself before connecting them to your agent.
For teams and organizations, Agensi's security scanning provides a meaningful layer of assurance that reduces the burden of individual auditing.
SKILL.md support
This is Agensi's unique angle. While Smithery and Glama focus exclusively on MCP servers, Agensi also hosts SKILL.md skills — files that teach AI agents how to do specific tasks. MCP servers give agents access to tools. SKILL.md skills give agents expertise. Both matter, and Agensi is the only platform that covers both.
If you want a code review skill that uses a GitHub MCP server to check PR history before reviewing code, Agensi is the only place you'll find both the skill and the server in one place.
Pricing
All three platforms let you browse and install free MCP servers at no cost. Agensi adds a marketplace dimension — some SKILL.md skills are paid, with an 80/20 revenue split favoring creators. Agensi Pro ($9/month) provides an MCP server endpoint that gives AI agents direct access to the Agensi catalog.
Which to use
Use Smithery when you're searching for a specific niche MCP server and want the broadest possible catalog to search through.
Use Glama when you want a clean, well-organized browsing experience for discovering MCP servers by category.
Use Agensi when you want security-scanned MCP servers and SKILL.md skills in one place, especially if you're building workflows that combine tool access with agent behavior. It's also the right choice if you want to sell your own skills or MCP servers.
In practice, most developers bookmark all three and search across them depending on the task. The MCP ecosystem is still early enough that no single platform has everything.
Find the right skill for your workflow
Browse our marketplace of AI agent skills, ready to install in seconds.
Browse SkillsRelated Articles
Composio Alternatives: Best AI Agent Integration Platforms (2026)
Looking for Composio alternatives? Comparing the best AI agent integration platforms — from MCP-native setups to curated marketplaces.
6 min read
The MCP Marketplace: Why AI Agents Need a Curated App Store
The MCP ecosystem has 10,000+ servers but no quality layer. Why AI agents need a curated marketplace, and how the app store model applies.
6 min read
MCP vs REST API: When to Use Each for AI Agents
MCP and REST APIs both connect AI agents to external services. When to use each, how they differ on discoverability and standardization, and when to combine them.
5 min read