MCP Server - No stepId found for tool_call_id call_XXXXXX

[swainsa1]

I have a in house MCP server which i am using to make a call , Tool is registered sucessfully and makes a call to a docker container where my MCP server is running

MCP Server configuration

myMcpServerName:
type: stdio
command: docker
args:
- run
- -i
- --rm
- -e
- TOKEN=${TOKEN}
- our_mcp_image:latest
env:
TOKEN: "XXXXXX"

LOG

WARNING: The requested image's platform (linux/amd64) does not match the detected host platform (linux/arm64/v8) and no specific platform was requested
2025-05-13T14:13:40.996Z debug: [MCP][User: 681cb798d8eb058f02ef1c48][XXX] Transport sending: {"method":"notifications/initialized","jsonrpc":"2.0"}
2025-05-13T14:13:40.997Z info: [MCP][User: 681cb798d8eb058f02ef1c48][XXX] Connection successfully established
2025-05-13T14:13:40.997Z debug: [MCP][User: 681cb798d8eb058f02ef1c48] Updated last activity timestamp: 2025-05-13T14:13:40.997Z
2025-05-13T14:13:40.998Z debug: [MCP][User: 681cb798d8eb058f02ef1c48][XXX] Transport sending: {"method":"tools/call","params":{"name":"XXX","arguments":{}},"jsonrpc":"2.0",... [truncated]
[05/13/25 14:13:41] INFO Processing request of type server.py:534
CallToolRequest
2025-05-13T14:13:42.159Z debug: [MCP][User: 681cb798d8eb058f02ef1c48] Updated last activity timestamp: 2025-05-13T14:13:42.159Z
2025-05-13T14:13:42.164Z error: [api/server/controllers/agents/client.js #sendCompletion] Operation aborted No stepId found for tool_call_id call_Cz4b9PwDX20qmiDkwRbzIZ2R
2025-05-13T14:13:42.164Z error: [api/server/controllers/agents/client.js #sendCompletion] Unhandled error type No stepId found for tool_call_id call_Cz4b9PwDX20qmiDkwRbzIZ2R

[swainsa1]

On Further investigation i am getting the same error with puppeter too , Its able to launch a browser but fails with the same error in the chat model

puppeteer:
  type: stdio
  command: npx
  args:
    - -y
    - "@modelcontextprotocol/server-puppeteer"
  timeout: 30000
  initTimeout: 10000
  env:
    NODE_ENV: "production"
  stderr: inherit

[swainsa1]

Only customization i have is turned of Steaming options manually from the code . Any help is really appreciated

[swainsa1]

It turns out the issue is with Stream only . When we disable Stream manually currently code does not support it gracefully , I have verified without it works with both Azure Open AI and Open AI with out any issue . We are good for now hence closing it
FYI @danny-avila