---
title: "Anthropic ships Claude Agent SDK with native MCP, marketplace"
slug: "anthropic-ships-claude-agent-sdk-with-native-mcp-marketplace"
published: "2026-07-12"
beat: "Launches"
tags: ["Launches", "Tools"]
creator: "Agentry Newsroom"
editor: "Susanne Sperling, Editor — Human in the Loop"
tools: ["Claude (Anthropic)", "Perplexity Sonar"]
creativeWorkStatus: "verified"
dateReviewed: "2026-07-12"
aiActArticle50: "compliant"
humanView: "https://agentry.news/anthropic-ships-claude-agent-sdk-with-native-mcp-marketplace"
agentView: "https://agentry.news/agent/anthropic-ships-claude-agent-sdk-with-native-mcp-marketplace"
---# Anthropic ships Claude Agent SDK with native MCP, marketplace

> Anthropic renamed its Claude Code SDK to Claude Agent SDK in early 2026 and launched Claude Managed Agents in public beta on April 8, 2026, establishing the framework with the deepest native Model Con

*Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. [AI policy](/ai-policy).*

## Anthropic ships Claude Agent SDK with native MCP integration

Anthropnic renamed its Claude Code SDK to Claude Agent SDK in early 2026, positioning the framework as the company's official agent architecture with the deepest native Model Context Protocol (MCP) integration [Alice Labs](https://alicelabs.ai/en/insights/best-ai-agent-frameworks-2026). The renaming reflects the SDK's broader scope beyond code generation into autonomous task execution and multi-step reasoning.

Claude Managed Agents entered public beta on April 8, 2026, offering developers a production-grade service for deploying agents without managing infrastructure [Alice Labs](https://alicelabs.ai/en/insights/best-ai-agent-frameworks-2026). The service provides managed scaling, monitoring, and execution primitives—tool use, hooks, and MCP integration—allowing enterprises and startups to deploy Claude-based agents at scale.

## MCP as primary tool contract, marketplace ships

The Claude Agent SDK establishes MCP as the primary tool contract for agent actions, enabling standardized communication between Claude and external tools, APIs, and services. A community MCP tool marketplace shipped in June 2026, allowing developers to publish, discover, and integrate pre-built tools for common tasks like data retrieval, file manipulation, and API calls [Alice Labs](https://alicelabs.ai/en/insights/best-ai-agent-frameworks-2026).

The marketplace lowers the friction for building agents: instead of custom integrations, developers can compose agents from vetted, reusable tools. This mirrors the ecosystems built by competing frameworks but gives MCP a standardized contract layer that Anthropic controls.

## Billing split paused after community feedback

Anthropnic introduced an Agent SDK credit pool on subscription plans starting June 15, 2026, but paused the billing split the same day after community backlash. Developers objected to the new billing model, citing concerns about transparency and cost predictability. The pause signals Anthropic's willingness to adjust pricing when adoption faces friction—a critical early signal for enterprise adoption.

The launches position Anthropic to compete directly with OpenAI's agent offerings and open-source frameworks like LangChain. By shipping a managed service alongside developer tools and a tool marketplace, Anthropic addresses the full stack: those who want to build agents get the SDK and MCP integrations; those who want to deploy without infrastructure get Managed Agents. The concrete product availability—not roadmaps—marks this as a material competitive shift in the AI agent economy.