MCP Servers
MCP servers give AI agents structured access to external tools and data — databases, APIs, file systems, code. The protocol is the bridge. The tool selection is the discipline.
Where to Start
New to MCP? Start with the concept page. Understand what MCP is, how the architecture works, and why every tool you load costs tokens.
Choosing tools? Go straight to the adoption radar. It has team-specific load lists, a 6-gate evaluation checklist, and a governance protocol for keeping your toolset lean.
| Model Context Protocol | What MCP is, how it works, architecture, token economics |
| Tool Selection Radar | Adoption radar, team profiles, decision checklist, governance |
| Server Catalog | Reference list of available MCP servers by category |
Context
- AI Tools — The full modality-stack matrix, all tool layers
- Agent Protocols — Where MCP fits the broader protocol stack
- Artificial Intelligence — The full AI platform layer
Questions
Which MCP server category creates the most value for an agent working on your most frequent task — and is that category what you've actually configured?
- At what MCP server count does context window management become the binding constraint?
- Which server would you drop first if your session budget was cut in half?