Skip to content

Add ToolPiper as a local app for GGUF models#2090

Open
BenRacicot wants to merge 1 commit into
huggingface:mainfrom
BenRacicot:toolpiper-local-app
Open

Add ToolPiper as a local app for GGUF models#2090
BenRacicot wants to merge 1 commit into
huggingface:mainfrom
BenRacicot:toolpiper-local-app

Conversation

@BenRacicot
Copy link
Copy Markdown

@BenRacicot BenRacicot commented Apr 10, 2026

Summary

  • Adds ToolPiper to the local apps dropdown on GGUF model pages
  • ToolPiper is a macOS AI toolkit that runs inference locally on Apple Silicon (llama.cpp, CoreML, MLX)
  • Users download and run models via natural language in the built-in chat — the snippet is simply Download {model.id}
  • macOS-only, same GGUF model filter as Ollama

Changes

  • packages/tasks/src/local-apps.ts — snippet function + LOCAL_APPS.toolpiper entry
  • packages/tasks/src/local-apps.spec.ts — 2 tests (GGUF shown, non-GGUF hidden)

How it looks

The dropdown shows ToolPiper with a macOS pill on any GGUF model page:

In ToolPiper chat, type
Download bartowski/Llama-3.2-3B-Instruct-GGUF

Links


Note

Low Risk
Low risk: adds a new LOCAL_APPS entry and simple snippet output gated behind existing GGUF detection, plus tests; no changes to core inference or data handling.

Overview
Adds ToolPiper to the local-apps list as a macOS-only option that appears on GGUF (llama.cpp-compatible) model pages.

Introduces a new snippet generator that instructs users to run Download {model.id} inside ToolPiper, and adds tests to verify ToolPiper is shown only for GGUF models and the snippet content is correct.

Reviewed by Cursor Bugbot for commit d702e37. Bugbot is set up for automated code reviews on this repo. Configure here.

ToolPiper is a macOS AI toolkit that runs inference locally on Apple Silicon.
Users can download and run GGUF models via natural language in the built-in chat.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant