Hey HN,<p>OpenAI released the Agents SDK yesterday, which is great because of its simplicity. I just added MCP support for it, which is currently available as a fork here: <a href="https://github.com/lastmile-ai/openai-agents-mcp" rel="nofollow">https://github.com/lastmile-ai/openai-agents-mcp</a> (and on pypi as the openai-agents-mcp package).<p>You can specify the names of MCP servers to give an Agent access to by
setting its `mcp_servers` property.<p>The Agent will then automatically aggregate tools from the MCP servers, as well as
any `tools` specified, and create a single extended list of tools. This means you can seamlessly use MCP servers, local tools, OpenAI-hosted tools, and other kinds of Agent SDK tools through a single unified syntax -- and have them interact in the same Agent run loop!<p>Everything else stays exactly the same.<p>```<p>agent = Agent(
name="MCP Assistant",<p><pre><code> instructions="You are a helpful assistant with access to MCP tools.",
tools=[your_other_tools], # Regular tool use for Agent SDK
mcp_servers=["fetch", "filesystem"] # Names of MCP servers from your config file (see below)</code></pre>
)<p>```<p>The servers are configured in an `mcp_agent.config.yaml` file, very similar to how they are configured for Claude Desktop:<p>```<p>$schema: "<a href="https://raw.githubusercontent.com/lastmile-ai/mcp-agent/main/schema/mcp-agent.config.schema.json" rel="nofollow">https://raw.githubusercontent.com/lastmile-ai/mcp-agent/main...</a>"<p>mcp:<p><pre><code> servers:
fetch:
command: "uvx"
args: ["mcp-server-fetch"]
filesystem:
command: "npx"
args: ["-y", "@modelcontextprotocol/server-filesystem", "."]
slack:
command: "npx"
args: ["-y", "@modelcontextprotocol/server-slack"]
</code></pre>
```<p>I have submitted an issue and PR into the openai-agents-python repo [2], and my plan is instead of a fork, I will create an extension package for MCP support (coming later today).<p>I was able to do this pretty quickly (got it working yesterday) because I've been building the mcp-agent library (<a href="https://github.com/lastmile-ai/mcp-agent" rel="nofollow">https://github.com/lastmile-ai/mcp-agent</a>), which makes MCP server aggregation/connection really easy. I did a Show HN about it a few weeks ago [3].<p>Wanted to share here to get community feedback on whether this is useful, which will help me decide if I should dedicate more time to it.<p>[1] - <a href="https://github.com/lastmile-ai/mcp-agent" rel="nofollow">https://github.com/lastmile-ai/mcp-agent</a><p>[2] - <a href="https://github.com/openai/openai-agents-python/issues/23" rel="nofollow">https://github.com/openai/openai-agents-python/issues/23</a><p>[3] - <a href="https://news.ycombinator.com/item?id=42867050">https://news.ycombinator.com/item?id=42867050</a>