Comment by Ahmet Özel on "Run AI Locally on Your Mac with Ollama: Chat, Tools, RAG and a Custom Model (Real Numbers from a Mac mini M4)"

CommentRun AI Locally on Your Mac with Ollama: Chat, Tools, RAG and a Custom Model (Real Numbers from a Mac mini M4)

Ahmet Özel

AI Engineer. Computer Vision, RAG and LLM agents.

The tool loop is the part I would adjust before adding the second function from your examples. The displayed snippet calls get_disk_free() for every tool call, regardless of the requested function name. That works for the single-tool demonstration, but a request for system memory would need a different dispatch path.

A small function-name-to-handler mapping, with arguments validated for the selected handler, would make that extension explicit. I would also append response.message once before iterating over its tool calls, then append each tool result separately. Otherwise a response containing multiple calls repeats the same assistant message in the conversation history.

RainyTech

AI, coding and tech explained clearly. Real tutorials with real numbers. New YouTube videos every week.

23h ago

Good catch, thank you! You're right: I shortened the loop for the article, and that version always calls get_disk_free() and re-appends the assistant message for every call. The code in the repo does it your way: a TOOLS dict mapping each name to its function, response.message appended once, then TOOLS[call.function.name](**call.function.arguments) for each call. It doesn't validate arguments yet, so I'll add that and fix the snippet in the article. The full version is tools.py in the GitHub repo linked at the top of the post.

Thanks for separating the article snippet from the repository implementation. My criticism applies to the displayed loop, and the mapping plus single assistant-message append in the full version already addresses those two issues.

For argument validation, I would reject unknown tool names and unexpected fields before invoking a handler, then return a structured tool error tied to that call. A regression with two different calls in one response would check dispatch, result ordering and conversation history together, while a malformed argument case checks that neither handler runs accidentally.

Search Hashnode