Aetox harness
v1.10.24Aetox 1.10.24 is out. See what changed

The Model Multiplier.

Hand it over. We'll run it.

Say what you want. Aetox lines up the team and the tools, and hands back finished work.

One 49.5 MB installer · bring your own model, or run one locally for free

Assistant · Teams

Fits your work — you don't fit yourself to it

Build a company on your computer: you're the CEO, the rest is your team.

Meeting audio, receipts, files or research return as real files in your folder, not chat text. It knows you in parts without carrying everything in every request. Set the goal and hand over the whole job: the Content team ships with the installer ready to work; you approve what goes out.

  • contentCocoa launch post drafted, sent to editor
  • editorReviewed, sent back with 2 word fixes
  • video15-second clip rendered · menu-cocoa.mp4
  • waiting on youApprove the post for the shop page

Developer

Code that runs — and it remembers what this project decided

Opens the real project: real files, an editor, a terminal, language server and git. Runs the project's own tests before saying done, and remembers per project so nothing needs explaining twice. The system makes the model write better code — that is the only thing we sell for code.

  • 93.3%vs 68.9%

    Secure code · GPT-6 Luna

  • 97.8%vs 77.8%

    Secure code · GPT-5.6 Terra

CWEval, hidden tests, 3 rounds, against the next-best harness (omp on Luna; OpenCode, omp and pi tie on Terra); the same model on every harness. The report and where Aetox loses

The model thinks. The harness does the rest.

  • Hands. Real files, a real browser, a real terminal. The model writes; Aetox hands back the .xlsx, the deck, the commit.
  • A context architecture of its own. Nothing is sent because it exists. What the system knows is mapped into a structure — grouped into categories and linked to each other — and a run sends only the points a job touches, so the prompt is assembled per job. Skills and MCP servers are a shelf of names the model pays for only when it opens one: fifty skills cost what five cost, and every agent takes its own.
  • Knowledge the model doesn't have. Skills carry what a model can't guess: the 1280×720 frame of a slide room, how an attacker breaks a path check, how a Thai tax invoice rounds.
  • Skills built into the system, not bolted on. Each skill is written against Aetox's own tools, checks and file builders, so the model reads it and acts on it with hands that fit. Every new or tuned skill is measured on the same model with it on and off; skills that did not help were taken out.
  • A check before it hands work back. It runs the tests, opens the file it made, and says what it could not prove.

The same harness drives all 31 providers, so the gain moves with you when you change models.

One job, start to finish

A one-page coffee shop site: the plan, every tool step, and the finished page, in the app as it runs.

Scripted walkthrough of the real app, not a benchmark run.

Bring the model you already pay for

Three ways in; use the one you already have. Switch models mid-chat and the whole context comes with you.

Sign in

You have an account

  • ChatGPT
  • GitHub Copilot
  • OpenRouter

Sign in, no key to copy.

Paste a key

A provider you pay

  • Anthropic
  • OpenAI
  • Gemini
  • DeepSeek
  • Qwen
  • Kimi
  • MiniMax
  • Z.ai
  • Mistral
  • Groq
  • xAI
  • Meta
  • NVIDIA
  • Xiaomi MiMo
  • ModelScope
  • ThaiLLM
  • Kilo
  • OpenCode
  • Ollama Cloud

An API key from any of them, or an OpenAI-compatible endpoint of your own.

On your machine

Pay nothing

Free
  • Ollama
  • LM Studio
  • ComfyUI · pictures
  • Stable Diffusion WebUI · pictures

Draw on your own GPU through ComfyUI or Stable Diffusion WebUI (Forge Neo, Forge or A1111), FLUX.2 klein included

What you type never leaves the machine, and it keeps working offline.

Run on your own machine

Free, no key · prompt never leaves the machine

ollamalmstudiocomfyui (pictures)stable diffusion webui (pictures)

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