# MCP vs CLI tradeoffs for AI agents > [!quote] > "A CLI has way less token usage… but those differences become less and less relevant. One area where I like MCP better is that the tool discoverability is better." Grafana's team (GCX CLI vs MCP server, benchmarked on o11y-bench) finds parity of outcome with opposite trade-offs; the deployment environment (remote vs local shell) usually decides. ## Links - [[3. Resources/Permanent Notes/Tool-Interface Choice Shapes Agent Architecture More Than Model Choice]] — applies-in; the CLI-versus-MCP tradeoff shows how interface constraints shape agent architecture. - [[3. Resources/Permanent Notes/Zet · CLIs trade tool discoverability for token efficiency]] — evidence-for; the benchmark pairs lower CLI token cost with stronger MCP tool discovery. - [[Lit · MCP protocol enables conversational WordPress publishing]] — a concrete MCP deployment whose value rests on the discoverability advantage named here - [[Lit · Fable 5 prompting — effort and memory]] — both locate the design lever in agent scaffolding (interface, effort) rather than the model itself - [[3. Resources/MOCs/AI Cost & Tokenomics MOC]] — applies-in: the MOC lists this note under Connected Literature Notes (synthesis 2026-07-21 — run 3). ## Source - [[3. Resources/Readwise/Articles/CLI and MCP Basics for AI Agents|CLI and MCP Basics for AI Agents — The Context Window #4]]