smrati

How Do I Add Local OpenAI-Compatible Models (Ollama, vLLM, LM Studio) to Pi?

Pi connects to any local OpenAI-compatible server by adding a providers entry to ~/.pi/agent/models.json — you give it a baseUrl (the server's /v1 endpoint), an api type (openai-completions for the chat-completions API), a throwaway apiKey, and a list of model ids. No extension, no OAuth; local servers ignore the key, it just has to be present so pi marks the model "available" in /model.

A simple view

  pi session                                  your machine
   ┌──────────────┐   models.json says:      ┌─────────────────────────────┐
   │              │   "ollama →              │ ollama serve    :11434 /v1  │

Code Is the Source of Truth

When answering questions about how software in the current workspace actually behaves (runtime, configuration, APIs, CLI, builds, tests, deployment, or infrastructure), treat the current repository artifacts as the primary source of truth.

Prioritize evidence roughly in this order:

Executable source and configuration
- Application source

Scripts

Multiple SSH Keys settings for different github account

create different public key

create different ssh key according the article Mac Set-Up Git

$ ssh-keygen -t rsa -C "your_email@youremail.com"

	{
	"$schema": "https://opencode.ai/config.json",
	"model": "<remote_provider>/<default-model>",
	"provider": {
	"llama_cpp": {
	"models": {
	"<local-gguf-model>.gguf": {
	"_launch": false,
	"name": "<local-gguf-model>.gguf",
	"limit": {

	import xmltodict
	import json
	import os # Import os module

	def read_xml_to_dict(file_path):
	"""
	Reads an XML file and converts it into a Python dictionary.

	Args:
	file_path (str): The path to the XML file.

	<!DOCTYPE html>
	<html lang="en">

	<head>
	<meta charset="UTF-8">
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
	<title>Advanced KG Prompt Generator</title>
	<style>
	:root {
	--primary: #667eea;

	from langchain_community.embeddings import OllamaEmbeddings
	import faiss
	import numpy as np

	# Initialize Ollama embeddings model
	embedding_model = (
	OllamaEmbeddings(model="llama3.2:3b")
	)

	# Sample sentences

	import chromadb

	from loguru import logger


	class ChromaHelper:
	def __init__(self, host, port):
	self.host = host
	self.port = port
	self.chroma_client = None

	import numpy as np
	import re
	from sklearn.metrics.pairwise import cosine_similarity


	# Load GloVe embeddings
	def load_glove_embeddings(glove_file):
	embeddings_index = {}
	with open(glove_file, 'r', encoding='utf-8') as f:
	for line in f:

	class wordle:
	def __init__(self, input_data_path):
	self.input_data_path = input_data_path

	def get_path(self):
	print(self.input_data_path)

	def get_words_by_len(self, word_length):
	"""
	filter words of specific length