control over responses with Streamable HTTP

[DinoChiesa]

I'd love to see an example or demonstration of this; or citation of documentation that explains what's happening.
I am missing something at the moment.

I am using fastmcp 2.10.4. main includes the following:

    mcp.run(
        transport="http", host="0.0.0.0", port=port, path="/mcp", stateless_http=True
    )

...and I run it via dotenv run python main.py.

Everything works, but I am not clear if I am taking advantage of streamable HTTP or not. As I understand the MCP spec allows batches of requests (defined here). And I suppose when using a chatbot/agent , the agent will decide if it wants to batch requests up or not. Q1: as an author of an MCP server that wants to support this, I don't need to do anything explicit in my code to make that happen. If I use FastMCP, it will just work. Is that right?

Assuming that is so, is there a way I can influence whether my tool sends back individual results or a batch of them?

I can see the client sends a batch of requests and the FastMCP Server responds with a batch of responses. But in some cases an individual responses may take some time to produce. Q2: Is there a way I can design my mcp.tool functions so that I can tell FastMCP , "return this request outside of a batch, right now"? Or is this one of those things I should just not worry about. Just let it go, let the framework handle it?