jpicklyk
Help Required
I really like the concept and the overall idea behind the task-orchestrator. I’d like to know if it can work with a self-hosted LLM instead of Claude. I have an OpenAI-compatible server URL available.
Additionally, is it possible to turn this into a Python CLI tool by configuring MCP? Or is there a simpler way to run it without relying on Claude?
Thanks in advance :)
Since this tool is an MCP, it can work with a self hosted LLM however my main focus is with Claude as that is what I like to use. Any LLM that supports the usage of MCP can make use of this.
The beautiful thing about MCP is that it is language agnostic. I package the server side within Docker so you don't need to be concerned with the fact that I have written this using Kotlin and SQLite. Your LLM will use the MCP protocol to self explore and use the tools exposed.
Please note that the main branch makes use of Claude specific features such as Subagents and Skills which is not a common concept across LLMs. The core MCP tools for breaking down work into various containers is still available (projects, features, tasks) but if you are not using Claude, you will miss out on these other features which are extremely powerful.
