100% LOCAL · WINDOWS · NO CLOUD

Your own AI.
On your machine.

Aphelion runs powerful language models entirely on your PC — chat, characters, stories and code, fully offline. No servers to wire up, no Ollama-on-top-of-everything. One app that auto-fits the best model to your GPU and just works.

Download for WindowsFree · 64-bit · ~120 MB View on GitHub

No account. No cloud.No telemetry. Ever.Open source.

Aphelion Model: llama3.1-8b-instruct ▾ ● Auto-Fit · 11.2 / 12 GB VRAM — ▢ ✕

Story · The Last Ember

YOU

Continue the story. The protagonist enters the ancient ruin and discovers a hidden chamber.

ARIA NYX

The air grows colder as you descend the worn stone steps, footsteps echoing in the silence. At the heart of the ruin a chamber opens — vast and circular, a pedestal cradling a faintly glowing crystal at its center.

Continue the story…

LIVE 100% OFFLINE · YOUR GPU · TOTAL PRIVACY Your data never leaves your machine 🔒

A whole AI workspace — in one window.

Not a chat box bolted onto a server. Aphelion is one app where characters, stories, dialogue trees, and an expert assistant live side by side — all powered by models running on your own machine.

LOCAL

Stays on your machine

Every prompt, model, and conversation runs and lives on your PC. Your data never leaves it — online or off.

ONE APP

Zero setup, not a stack

Install one program and start talking. No servers, no containers, no Ollama-plus-six-other-things to wire together.

AUTO-FIT

Fits your hardware

Aphelion reads your GPU and VRAM and loads the best model that runs fast — no config files, no guesswork.

WORKSPACE

More than a chat box

Characters and group chats, long-form stories, branching dialogue trees, and expert assistants — all in one place.

PRIVATE

Private by default

No account. No telemetry. No phone-home. The lock icon means exactly what it says.

OPEN

Open and yours

Free and MIT-licensed. Read the source, change it, build on it. Nothing hidden.

It picks the right brain for your machine.

On first run, Aphelion checks your graphics card and downloads a model sized to fit — so it runs fast instead of crawling. Bigger card, bigger and smarter model. Switch any time, or bring your own.

~6 GB VRAM

Light & fast

A compact model like Gemma 3 4B. Runs on modest GPUs — even CPU in a pinch.

~2.3 GB download

~12 GB VRAM

Balanced

A strong all-rounder like Gemma 3 12B for mid-range cards.

~7 GB download

20 GB+ VRAM

Most capable

Flagship models — Gemma 3 27B or SuperGemma4 26B — for high-VRAM GPUs.

~15 GB download

Uncensored models are available behind an opt-in warning, and you can point Aphelion at your own GGUF file. Whatever you pick, it runs entirely on your machine.

It's a model, not an agent.

Some "AI" tools are autonomous agents — deliberately given permission to run commands, browse the web, and change files on your computer (think AutoGPT, OpenClaw, and computer-use bots). That power is also the risk. Aphelion is the opposite: a language model in a box that only speaks text.

WHAT IT DOES

Reads your prompt and writes text back — that's the whole job
Runs a model on your GPU, on your machine
Saves the chats and characters you make, on your own disk

WHAT IT CANNOT DO

Run programs, scripts, or shell commands
Read, change, or delete your files
Reach the internet, your accounts, or other apps
Click, type, or control anything on your computer

There are no system permissions to grant and nothing to sandbox — because the model was never given any. It can write a villain's monologue; it can't go rummaging through your hard drive.

Questions, answered.

Is it really free?

Yes — completely. MIT open source, no account, no trial, no upsell.

Does it send my conversations anywhere?

No, never. The AI runs on your own computer. Disconnect from the internet and it keeps working.

Is this an "AI agent"? Can it do things to my computer?

No. Aphelion only generates text inside the app — it can't run commands, read or change your files, or touch anything else on your system. Unlike autonomous agents, it has no permissions at all. See "It's a model, not an agent" above.

What's a "model"?

The AI's "brain" — a file the app downloads once. Setup picks a good one for your hardware; you can change it in Settings.

Windows says it "protected my PC" — is that bad?

No — it just means the app isn't code-signed yet, and that certificate is a recurring monthly cost, so while Aphelion is free I haven't bought one. Click More info → Run anyway (the source is all public). If enough folks chip in, I can sign it and the warning goes away.

What hardware do I need?

64-bit Windows 10 or 11. A graphics card (GPU) is recommended — more VRAM means a bigger, smarter model — but setup fits a model to whatever you have.

Can I use uncensored models?

Yes — behind an opt-in warning. Whatever the AI writes is generated by the model you choose, not by Aphelion. Use at your own discretion.

How do I uninstall it?

Settings → Apps → Aphelion → Uninstall. It even asks whether to also delete the downloaded model and your chats, so you can reclaim the disk space.

Free — and staying that way.

Aphelion is free and open source. It's also unsigned — the only reason Windows shows that "unknown publisher" warning — because a code-signing certificate is a recurring monthly cost, and I haven't paid for one. If donations start coming in, that's the first thing they'll go toward. They also help keep Aphelion free: the more they offset what it costs to build and ship, the less chance I'd ever have to charge for it. No pressure — if it's useful to you, anything helps.

No pressure, ever — the app stays free either way. Donations just decide how fast the rough edges (like that signing warning) get smoothed out.