v0.1.3 · MIT Licensed · Open Source

The package manager for local AI models¶

ModelDock discovers, downloads, caches, verifies, and loads local LLMs through pluggable runtime adapters. No more manual ollama pull — just md.load("llama3") and it just works.

Get Started View on GitHub

ModelDock demo: installing modeldock, loading a model with the Python SDK, and browsing models with the CLI

Get Started¶

Quickstart Get ModelDock running in 30 seconds. Get started →

Installation pip install, optional backend extras, and requirements. Get started →

First Model Load, chat with, and manage your first local model. Get started →

Core Features¶

Python-First API One line to load a model — md.load("llama3"). No terminal commands, no repetitive downloads. Learn more →

Auto-Install If Missing load() checks, downloads, verifies, and returns a ready-to-use client. Learn more →

Searchable Registry Browse models, categories, capabilities, and sizes without leaving Python. Learn more →

Bulk Installation Install entire categories of models at once — md.install_category("coding"). Learn more →

Smart Caching Never re-download installed models. Content-addressed offline cache. Learn more →

Extensible Runtimes Ollama, LM Studio, and llama.cpp ship today; more are drop-in adapters. Learn more →

Quick Example¶

Python SDKCLI

import modeldock as md

# Auto-install if missing, then use
client = md.load("llama3")

# Run a chat
response = client.chat(
    model="llama3",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response)

# Load a model (auto-installs if missing)
modeldock load llama3

# Browse the catalog
modeldock list

# Search by capability
modeldock search coding

# Install specific models
modeldock install llama3 qwen3 deepseek-r1

Supported Runtimes¶

Runtime	Status	Notes
Ollama	Shipped	First runtime, fully supported
LM Studio	Shipped	OpenAI-compatible local server adapter
llama.cpp	Shipped	`llama-server` OpenAI-compatible adapter
Jan AI	Planned	Drop-in adapter
GPT4All	Planned	Drop-in adapter
vLLM	Planned	Drop-in adapter

Resources¶

Architecture Clean Architecture design, ports/adapters, and how it all fits together. Read the docs →

Contributing Guidelines, dev setup, and how to add a new runtime adapter. Get involved →

GitHub Source code, issues, and 100+ good-first-issues welcome. View repo →

PyPI Package pip install modeldock — releases, versions, and downloads. View package →

Changelog What shipped, what changed, and what's next. See what's new →

Security Supported versions and how to report a vulnerability. Read more →

Install¶

pip install modeldock
# with the Ollama backend helper (optional):
pip install modeldock[ollama]

Requires Python 3.9+ and a local Ollama install for the first runtime.

ModelDock is created and maintained by Himanshu kumar (OpenAgentHQ).

Licensed under MIT.