MCP Integration unhandled errors in a TaskGroup (1 sub-exception)

ChatGPT Apps SDK
1

Greetings,

i tried to use the local connectors of ChatGPT desktop
For this a createt an MCP Server on localhost.

The Problem is everytime i try to add the server it gives me the error
”unhandled errors in a TaskGroup (1 sub-exception)”

image
image654×1054 31.7 KB

From what i see in my logs he is not even trying to connect to that port.

Does anone had simlar issues?

2

Update after hours of debugging i think it is a in beta a cloud service and can’t connect to localhost, i never saw a try .. not even to do a SSL handshake.

3

Can someone of OpenAI confirm that there are no Debug Tools? Or that it is curently not possible to use Windows ChatGPT against a Localhost MCP Server? I’m asking because there are rumors you can use it directly on localhost but i can’t confirm this.

4

Hey all, I am also running mcp in localhost and a static ngrok domain. Keep getting the TaskGroup (1 sub-exception) error. Has anyone managed to solve this already?

5

Sorry i currently wasn’t able to resolve it, from my understanding it it does not currently work on localhost adresses. But soon i got an update i will tell you

6

+1 using ngrok domain https://nebuly-unfervid``…

It was working fine 2 days ago but stoped working yesterday (Oct 16). Here the full response object:

{

“type”:“http_error”,

“code”:424,

“message”:“424: unhandled errors in a TaskGroup (1 sub-exception)”,

“is_error”:true

}
7

I also deployed the mcp to railway (nothing is on localhost) and still get the same error. I guess it’s indeed an error in OpenAI.

8

Thank you for the information

9

I was getting this error because I was returning wrong uri in registerResource (where you set the mimeType: "text/html+skybridge")

And because ChatGPT calls resources/list when adding a connector, it failed.
But I agree that the error message could be less cryptic

10

did anybody find a solution?

I am using a mcp server, not local server though, still getting Error refreshing actions.

unhandled errors in a TaskGroup (1 sub-exception), and tried all kinds of ways claude code and codex suggested.

11

I was able to solve this problem, so here’s a solution for posterity.

I hit the same generic error when calling tools through ChatGPT:

ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)

The MCP server itself was healthy, the Secure MCP Tunnel was connected, and ChatGPT could discover the connector and its tools. However, every tool call failed. ChatGPT sometimes presented this as an expired connection, which made it initially look like an authentication or tunnel problem.

The issue was MCP resource discovery. ChatGPT calls resources/list while initializing or using a connector. My server registered tools but did not register any resources. Adding one valid, readable resource fixed every tool call immediately:

server.registerResource(
  "campaign-tracker-about",
  "campaign://tracker/about",
  {
    title: "Campaign tracker information",
    description: "Identifies this MCP server and its purpose.",
    mimeType: "text/plain",
  },
  async (uri) => ({
    contents: [{
      uri: uri.href,
      mimeType: "text/plain",
      text: "Live campaign tracker.",
    }],
  }),
);

The important details were:

  • The resource had a valid URI: campaign://tracker/about.
  • The URI returned by the read callback matched the registered URI.
  • The advertised and returned MIME types matched.
  • Both resources/list and resources/read were tested directly afterward.

I also added this regression test:

const resources = await client.listResources();

assert.deepEqual(
  resources.resources.map((resource) => resource.uri),
  ["campaign://tracker/about"],
);

const about = await client.readResource({
  uri: "campaign://tracker/about",
});

assert.match(about.contents[0].text, /Live campaign tracker/);

After restarting the MCP server and refreshing the ChatGPT connector, get_current_location and the other tools began working through the same tunnel and authentication configuration.

Environment:

Node.js v25.8.0
@modelcontextprotocol/server 2.0.0
@modelcontextprotocol/node 2.0.0
@modelcontextprotocol/client 2.0.0
OpenAI Secure MCP Tunnel

I cannot see the nested exception that ChatGPT swallowed, so I cannot prove the precise internal cause. However, adding a valid resource catalog was the only material server-side change, and it changed the connector from consistently failing to consistently working. A more specific error mentioning resources/list or resource registration would have made this substantially easier to diagnose.