No content response received.

[JCLiu-hfut]

When an agent invokes a tool, sometimes the text is suddenly truncated. It is speculated that this may be because the large model has reached the maximum context window？

[jonigl]

Hey @JCLiu-hfut thanks for reaching out!

It could be realted to the context window. However, to help narrow it down, could you please share a bit more context about your setup?

• ollmcp version
• Ollama version
• Operating system
• Model name + variant (e.g., gpt-oss:20b-cloud, or a local model?)
• What MCP Servers do you have configured?
  • How many tools do you have enabled?

If you can share that info, I’ll be happy to help debug further 🙂

[JCLiu-hfut]

Thank you for your reply. My configuration at the time was as follows:

ollmcp version: 0.22

ollama version: 0.12.10

Operating system: macOS

Model: qwen3:8b (local)

MCP: An MCP service containing 22 tools

[jonigl]

Hey @JCLiu-hfut, thanks for sharing your setup!

I have a couple of follow-up questions:

1. Are you using Thinking Modewith qwen3:8b (local)?
2. Could you share the MCP servers you have configured? That way I can try to replicate what’s happening on my side.

In the meantime, one suggestion is to disable any tools you don’t need for your query. This can reduce the number of tokens in the context window and might help prevent truncation. If you get a chance, please test this and let me know how it goes.

Thanks, and looking forward to your reply!

[JCLiu-hfut]

Hello, thank you for your reply. I am using the locally running Thinking Mode qwen3:8b version, and the MCP service I am using comes from this project: https://github.com/haris-musa/excel-mcp-server. Ultimately, I only loaded this one MCP service.