Open Source
The LLM is the runtime.
Every system call, every decision, every output — all LLM.
openall is the open source foundation
for what comes after code.
The Shift
The code between a user and an LLM was never architecturally necessary. It existed because models weren't yet capable of being the runtime themselves. That world is over. The stack collapses. What remains is the only two things that were ever real: intent and result.
Before openall
With openall
That's the entire stack.
Artificial Super Assistance
openall turns your device into an ASA — an Artificial Super Assistant that knows your context, remembers everything, controls your environment, and delivers exactly what you need, the moment you need it. No downloads. No configuration. Just intent, and a persistent tool that lives on your machine.
01
Personal Productivity
"I need a daily command center — my meetings, my files, my emails, my calendar. Let me book a restaurant or move a file from the same place."
What ASA delivers
A persistent, windowed tool that surfaces everything relevant to your day. It knows your schedule, your open threads, your preferences. It acts — books, drafts, moves, reminds — without switching apps or losing context.
02
Professional Workflow
"Give me a tool to run my weekly review — pull the data, update the deck, flag what changed, and send the summary to the team."
What ASA delivers
A workflow instrument that runs on your schedule. It reads your data sources, rewrites the relevant slides, surfaces anomalies, and drafts the message — all in one session, with full memory of last week's run.
03
The Impossible Demo
"I'm closing a deal. I need a room for it."
What ASA delivers
A bespoke deal room that has never existed before and will never exist again. A windowed interface built around this relationship, this timeline, this person's way of thinking. Not a dashboard pulled from a template. Not a notification pipeline. A living instrument: the emails on one side, the open questions on the other, the silence tracker in the corner that knows this stakeholder always goes quiet before they say no. It remembers last Tuesday. It updates as the deal moves. When it's done, you close it. It was always only ever yours.
Every one of these tools was described in plain language, delivered instantly, and runs where you need it — with persistent memory and a native windowed interface. No code was written at any point.
How it works
01 / Install
A native installer for Mac or Windows. No dependencies, no configuration, no dev environment to set up. openall runs entirely in your environment — nothing leaves without your say.
02 / Connect
Paste in your API key for the LLM provider of your choice. openall connects directly to the model. No accounts, no subscriptions, no middleman. Your key, your model, your control.
03 / Describe
No syntax. No boilerplate. No scaffolding. Describe your intent in plain language. A tool. A workflow. A data transformation. A product. Anything.
04 / Done
Your chosen model generates, executes, and evolves the result in real time. No code is written. No file is saved. The model is the application.
In action
A user. A sentence. A running result. No code written at any point.
01 — The platform is installed and ready for use
02 — The users asks in plain language.
03 — openall runs it. No code written at any point.
What you get
No telemetry. No data harvesting. No intermediary between you and your model. Your intent, your context, and your output stay yours, wherever you run it.
Native installers for both platforms. Download, install, and you are running in under two minutes. No dependencies required.
The full codebase is on GitHub. Read it, fork it, extend it, or contribute to what comes next. Built in the open, for the open.
Not low-code. Not AI-assisted code. Not vibe coding. Zero application code, by design and by architecture. This is not a tool on top of the old world. It is the end of it.
Install openall
Open source. Ready now.