← Retour au blog
tech 5 August 2026

Stateless MCP: Why It's Making a Comeback

The Stateless Model Context Protocol (MCP) is back with its 2.0 version, bringing significant improvements for agent frameworks. Learn why this protocol is regaining popularity.

Article inspired by the original source
Stateless MCP has recaptured my interest ↗ simonwillison.net

Introduction

There are times when a forgotten technology resurfaces, revitalized by significant improvements that rekindle the interest of developers and tech decision-makers. This is precisely what's happening with the Stateless Model Context Protocol (MCP). The 2.0 version of MCP, launched on July 28, 2026, brings a fresh breath of air to this protocol that had been overshadowed by other innovations.

What is Stateless MCP?

The Model Context Protocol is a standard for exposing new tools to agent frameworks powered by language models (LLM). The main change introduced by version 2.0 is its stateless nature, greatly simplifying interactions between clients and servers. Previously, MCP required multiple HTTP requests to initiate a session and call a tool. Now, a single request suffices, reducing the complexity of development and maintenance.

Why MCP Had Lost Ground

Introduced by Anthropic in November 2024, MCP saw a peak of interest in 2025 before giving way to more flexible solutions like Skills, capable of interacting with a terminal environment. However, this flexibility came with risks, particularly in terms of security and control.

Advantages of Stateless MCP

The simplicity of stateless MCP makes it particularly attractive. MCP tools are easier to audit and control, a crucial asset in a context where data security is paramount. Moreover, the lightweight nature of the protocol allows smaller models, running on devices like laptops, to use them effectively.

Implementation Example

Let's take a concrete example to illustrate the difference. The old MCP required two requests: ``plaintext POST /mcp HTTP/1.1 Content-Type: application/json { "jsonrpc": "2.0", "id": 1, "method": "initialize", "params": { "protocolVersion": "2025-11-25", "capabilities": { }, "clientInfo": { "name": "my-app", "version": "1.0" } } } ``

With stateless MCP, only one request is necessary: ``plaintext POST /mcp HTTP/1.1 MCP-Protocol-Version: 2026-07-28 Mcp-Method: tools/call Mcp-Name: search Content-Type: application/json { "jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": { "name": "search", "arguments": { "q": "otters" }, "_meta": { "io.modelcontextprotocol/clientInfo": { "name": "my-app", "version": "1.0" } } } } ``

Implications for the Future

The comeback of stateless MCP opens the door to new possibilities for developers looking to integrate language models into their applications in a secure and efficient manner. By minimizing the complexities associated with state management, stateless MCP allows efforts to be focused on innovation and optimization.

Conclusion

Stateless MCP is an elegant solution to the challenges posed by modern agent frameworks. Its simplicity and security make it a wise choice for companies looking to leverage LLMs without the risks associated with more permissive environments.

Let's discuss your project in 15 minutes.

Stateless MCP Model Context Protocol LLM Anthropic Agent frameworks
Deepthix newsletter · 100% AI · every Monday 8am

An AI agent reads tech for you.

Our AI agent scans ~200 sources per week and ships the best articles to your inbox Monday 8am. Free. One click to unsubscribe.

Visit the newsletter page →

Want to automate your operations?

Let's talk about your project in 15 minutes.

Book a call