Aetoxv1.9.3 · Windows

Aetox

Pricing

The program is free. You only choose who does the thinking — a model on your own machine, or a provider you already pay.

Free. Choose who does the thinking

The program costs nothing — no subscription of ours, no paid tier, no account to sign up for. You only choose which model thinks for you, and you can switch mid-conversation.

Run it on your own machine

Freeno key needed

  • Connects to the Ollama or LM Studio you already have
  • What you type never leaves the machine
  • Keeps working if the internet drops

Best for everyday work · data that must not leave the organisation

OllamaLM Studio

Use a provider you trust

Pay as you go27 providers

  • You pay them for what you use. We add no markup
  • Switch provider or model mid-conversation — the context follows
  • Your machine talks to them directly, never through us

Best for hard jobs · work that needs the largest models

AnthropicOpenAIGeminiDeepSeekCopilot+19

No need to decide yet. The app ships with a demo model you can play with right after install — see the real screen, watch it pick up tools, read its full reasoning, without connecting anything.

Light enough to leave open all day

An assistant you close because it slows your machine is one you do not use. Aetox is two files, 90.4 MB together — the window and its engine — drawn with the WebView2 that Windows already ships.

Smaller than Claude Code 3× · Cursor 10× · VS Code 13×. Different categories, of course — a CLI has no window and an IDE is a tool for a different job; the fair comparison is the disk each takes on a machine like yours. Competitors measured 27 Jul 2026, Aetox 27 Sep 2026 on v1.9.0, both files — method in BENCHMARK.md.

And while it works

Task Manager on a real afternoon: the Aetox process at 64 MB while an IDE next to it sat at 1.7 GB. Straight talk, as before — Task Manager files the WebView2 that draws our window under Microsoft Edge WebView2, not under Aetox, so add that in your head. What wins is that you are not keeping a second copy of a browser.

90.4MB
Disk on your machine
Two files · installer 36.7 MB
5,754tests
Pass before release
Go 3,648 · UI 2,106
0.32ms
One turn assembled
174.9 KB memory · 13 Aug 2026
35tools
In the model's hands from install
About 10,700 tokens per request

Run on your own machine

Free, no key · prompt never leaves the machine

ollamalmstudio

Use a provider you trust

Enter a key for one you already have, or log in with GitHub Copilot · ChatGPT · OpenRouter · talk straight from your machine to them, not through us.

github copilotkiloanthropicopenaigeminideepseekkimiminimaxopenroutermistralqwengroqzaixaithaillmmodelscopenvidiaollama cloudchatgpt (codex)opencode zenopencode gometaxiaomi mimoxiaomi mimo token plan

Click a name to see each provider's details · plus any OpenAI-compatible endpoint of your own · switch at any time, no lock-in