PaglaAI

PAGLA · LOCAL AI, OPENLY BUILT

The local AI ecosystem

PaglaAI is a family of local-first tools — models, gateways, and launchers that run on your hardware and speak to each other openly.

Explore products Read the docs

One system across many surfaces

Every PaglaAI product inherits the same mark, type, and values. A design authority keeps the ecosystem consistent, and a shared launcher keeps it local. You never leave your machine to talk to it.

The ecosystem

  • PaglaMLX — LLM orchestration optimized for Apple Silicon, managed from the menu bar.
  • PaglaROUTER — a Cloudflare Worker gateway routing requests across providers.
  • PaglaDESIGN — the design authority: tokens, components, and brand, versioned.
  • PaglaCPP · PaglaAPI · PaglaGPT · PaglaOS — the roadmap beyond, in the same family.

What exists today

PaglaMLX

A local gateway for MLX models and cloud providers, in one menu-bar app.

Learn more

PaglaROUTER

A serverless gateway that routes LLM traffic across providers with failover.

View source

PaglaDESIGN

The design system every surface inherits — tokens, components, and brand.

View source

How it is built

Local first

Your models, your data, your hardware. The cloud is a fallback, not a requirement.

Open, openly

Source is public, decisions are recorded, and the roadmap is on the page.

Simple and precise

Whitespace is the loudest element. Design communicates before it decorates.

Start here

bash 
$ python setup.bat
$ python paglaai_core.py
# Gemini, Kilo Code, Claude Code, Qwen Code — one launcher.

Getting started

The documentation moves from beginner to advanced — install, quick start, concepts, guides, and the API reference.

Open the docs

Latest updates

Built in the open, for your machine

Follow the work, open an issue, or run it yourself. The ecosystem is one checkout away.

View the organization