Anthropic · The Agentic Era
Model Context Protocol (MCP)
The open protocol that standardised tool connections for LLM agents. Now an industry standard.
By C.W. Jameson · Published 19 May 2026 · Last reviewed 19 May 2026
MCP (Model Context Protocol) was published by Anthropic in November 2024 as an open standard for how LLMs connect to external tools, databases, and services. Within six months, every major IDE extension and several frontier model providers had adopted it. The significance is architectural: MCP made tool definitions portable across models in the same way that USB made peripherals portable across computers.
Field signature
JSON-RPC protocol for tool invocations; discoverable tool manifests.
Specifications
| Released | 2024-11 |
|---|---|
| Context window | N/A (protocol layer) |
| Pricing | Free (open protocol) |
| Modalities | tools · resources · prompts |
| License | MIT |
| Era | The Agentic Era |
Strengths
- Portability across models
- Ecosystem adoption
- Open standard
Weaknesses
- Authentication complexity
- Security review still maturing
Authentication markers
The fingerprints by which Model Context Protocol (MCP) can be identified from its output alone.
| Tell | Meaning |
|---|---|
| JSON-RPC tool calls with a manifest discovery phase. | MCP-native agent. |
Notable works
- Became the cross-vendor tool-use standard within six months of release
Market position
Free (open protocol)
Partner offer
Anthropic's Claude family is the model lineage most operators end up on for serious agent work. The free tier remains useful.
Try Claude →Affiliate link — see disclosure.
Primary sources
- [1] Anthropic: MCP
From the Almanac shop
The Operator's Compendium
Every agent harness, every routing pattern, every cost trick. 90-page PDF.
$29 — Coming soon