ezollama 0.2.4

A lightweight Python wrapper for Ollama AI models.

EZollama

A simple Python library for interacting with Ollama models via their local API and cloud models through Google, Groq, OpenAI and Anthropic.
Supports model selection, chatting, persistent system prompts, listing models, downloading models, resetting chat history, and text-to-speech.

---

Installation

1. Install Ollama:
   Download and install Ollama from https://ollama.com/download.
   The library will prompt to install Ollama if not found.

2. Python dependencies:
   The library auto-installs pyttsx3 for text-to-speech if missing.

---

Usage for local

Import

from ezollama import EzOllama
ez = EzOllama()

Set Model

ez.set_model("llama2")

Set Persistent System Prompt

ez.set_system_prompt("You are a helpful assistant.")

Chat

response = ez.chat("Hello!")
print(response)

List Available Models

models = ez.list_models()
print(models)

Pull (Download) a Model

ez.pull_model("llama2")

Reset Chat History

ez.reset_history()

Text-to-Speech

ez.text_to_speech("Hello, this is AI speaking.")

---

Usage for cloud

Import

from ezollama import EzOllama
ez = EzOllama()

Set mode

ez.set_mode("mode", "api-key") # choosable from groq, google, anthropic and openai

Set model

ez.set_model("model") # for example 'gemini-2.5-flash

Set Persistent System Prompt

ez.set_system_prompt("You are a helpful assistant.")

---

Example for local

from ezollama import EzOllama

ez = EzOllama()

ez.set_model("llama3.2:3b")
ez.set_system_prompt("You are a friendly assistant.")

while True:
    user_input = input("- ")
    resp = ez.chat(user_input)
    print(resp)
    ez.text_to_speech(resp)

---

Example for cloud

from ezollama import EzOllama

ez = EzOllama()

ez.set_mode("google" "API-KEY")
ez.set_model("gemini-2.5-flash")
ez.set_system_prompt("You are a friendly assistant.")

while True:
    user_input = input("- ")
    resp = ez.chat(user_input)
    print(resp)

Notes

• The library checks and quietly starts the Ollama server before each API call.
• If Ollama is not installed, you will be prompted to install it.
• If the model does not exist, pull_model will print a message.
• Text-to-speech uses pyttsx3 and works cross-platform.