Best MCP / Agent Infrastructure Tools

Arcade.dev gives developers a secure MCP runtime for AI agents, including user authorization, hosted tools, and production controls.

Key Features:

• •Chrome extension screen capture for instant demo recording
• •Hotspots, guided tooltips, and callout annotations
• •Embeddable demos with customizable player appearance

IBM-originated open-source document processing software for parsing, understanding, serializing, and chunking complex documents for AI pipelines.

Key Features:

• •Document Format Conversion
• •Layout Analysis and Reading Order
• •Table Structure Recognition

the open protocol specification and documentation site for connecting AI applications with tools, resources, prompts, and data systems.

Key Features:

• •Open-source standard for connecting AI applications to external systems
• •Client/server architecture for tools, resources, prompts, and data sources
• •Local and remote MCP server support with versioned specification docs

Smithery is an MCP marketplace and infrastructure layer for connecting AI agents to services. The homepage fetched during this run described the product as “Connect agents to services in minutes” and “Give agents more agency,” with auth, credentials, and sessions handled for users. It also showed 715 MCPs, which is a useful concrete signal: Smithery is not just a single connector, but a discovery surface for many Mod

Key Features:

• •Public catalog link advertising 9,538+ MCPs at research time
• •Searchable MCP registry with listings such as Context7, Microsoft Learn MCP, Browserbase, Airtable, and Blockscout MCP Server
• •CLI install and auth flow using npx smithery auth login, npx smithery mcp add, and tool-call commands