explaingit

merken/pi-local-models

Analysis updated 2026-05-18

★ 1TypeScriptAudience · developerComplexity · 2/5Setup · moderate

TLDR

A PI Coding Agent add-on that discovers and connects to AI models running locally on your own machine.

Mindmap

mindmap
  root((pi-local-models))
    What it does
      Discovers local AI models
      Works with OpenAI style APIs
      Plugs into PI Coding Agent
    Tech stack
      TypeScript
      llama.cpp
      Ollama
      oMLX
    Use cases
      Use local models in PI
      Switch model servers
      Per project configs
    Audience
      Developers
      PI Coding Agent users
      Local AI hobbyists

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Use a locally running AI model like Ollama or llama.cpp inside the PI Coding Agent.

USE CASE 2

Switch between different local model servers using PI's /models command.

USE CASE 3

Keep separate local model configs for different projects.

What is it built with?

TypeScriptOpenAI API specllama.cppOllama

How does it compare?

merken/pi-local-models0xkinno/neuralvault0xmayurrr/ai-contractauditor
Stars111
LanguageTypeScriptTypeScriptTypeScript
Setup difficultymoderatehardeasy
Complexity2/54/52/5
Audiencedeveloperdeveloperdeveloper

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 30min

Requires already having a local model server like Ollama or llama.cpp running.

In plain English

pi-local-models is a small add-on for the PI Coding Agent that helps it find and use AI models running locally on your own computer instead of only relying on cloud based models. It works with any local model server that speaks the same API style as OpenAI, which covers several popular local tools such as llama.cpp, Ollama, oMLX, llama-swap, and MTPLX. Once installed through PI's own package installer, you set it up by creating a small configuration file that points to your local model server, giving it the web address it is running at and an API key. This config file can be placed either for your whole user account or just for a single project, so you can have different local model setups for different projects if you want. After the configuration is in place, you run PI and use its built in /models command, which will then discover and list any local models it finds through this tool, letting you pick one to use inside PI. The readme is intentionally brief. It does not describe internal architecture, testing, or license terms, and it points to a separate example config file in the repository for a fuller reference.

Copy-paste prompts

Prompt 1
Help me write a local-models.json config file to connect PI to my Ollama server.
Prompt 2
Explain how to install pi-local-models and set it up at the project level.
Prompt 3
What local model servers work with this tool besides Ollama and llama.cpp?
Prompt 4
Walk me through using the /models command in PI after installing this.

Frequently asked questions

What is pi-local-models?

A PI Coding Agent add-on that discovers and connects to AI models running locally on your own machine.

What language is pi-local-models written in?

Mainly TypeScript. The stack also includes TypeScript, OpenAI API spec, llama.cpp.

How hard is pi-local-models to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is pi-local-models for?

Mainly developer.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.

Alternatives & similar repos