Documentation

Autonoma is a Windows desktop application. It runs on your machine, uses your accounts, and calls AI and data providers with API keys you supply. Nothing here requires a server of ours to be up.

These pages describe what the software does today. Where something is switched off by default or not implemented, that is stated rather than glossed over.

Start here

  1. Install and licence — get it running and entered.
  2. API keys — the app ships with none, so this is the step that makes anything work.
  3. Social Studio — connect an account and plan your first posts.

The shape of the thing

Autonoma is fifty-eight screens in one window, grouped into eleven areas. You are not expected to use all of them — pick the two or three areas that match your work and ignore the rest. Anything you have not configured reports itself as unconfigured rather than pretending.

AreaWhat is in it
DashboardOverview with system scan and bot controls, and an agentic CLI that can drive the rest
AnalyticsRevenue and activity charts (Studio plan), revenue by stream and platform with CSV export (every plan)
Income and publishingStorefront with Ebook Studio, audiobook builder, a Library with real PDF/EPUB/audio readers, passive-income channels, creator marketplace and partner programme
Trading and cryptoTrading, wallet, per-chain balances, mining, claiming, token deployment — all off by default. Moving funds between chains is not implemented
AutomationBots, connected platforms, pricing rules, campaigns, pipelines
AI and evolutionProvider keys, text/image/audio/video generation and MediaStudio Pro, a self-improvement loop, DAO governance (Snapshot is read-only; proposals and votes you create are local), the JK2 agent, and the game-adapter layer behind it
DevelopmentA full IDE with LSP and debugging, VPS management over SSH, Jedi Knight modding tools
Reverse engineeringStatic analysis of executables, firmware, WebAssembly, Android and .NET
CommunicationDiscord bot and announcements, social accounts and the Social Studio
SystemBackground service, feature toggles, security centre, settings, logs
StudioThe workings underneath: workspace layout, plugins, telemetry, data contracts, legacy migration and operations — including the things each one refuses to do

This page used to publish a total of forty-six, on the same day the features page derived forty-five from its own list. Both were wrong, the two disagreed with each other in public, and neither of them included the Studio area at all. Every count on this site is now checked against the app’s own navigation table on each build.

A full screen-by-screen breakdown, with the known limitations of each, is on the features page.

The crypto features are governed by explicit switches. Storing an exchange API key does not enable trading — the credential and the permission are deliberately separate decisions. See safety and dry run.

Where your data lives

Your data is under your user profile, not in the installation directory:

%USERPROFILE%\.autonoma\
  data\credentials\     your API keys, encrypted
  data\social\          connected accounts and the content plan
  data\safety\          the dry-run switch and its journal
  data\security\        the safety state and the STOP file

Those last two are different folders on purpose and this page used to name only one, saying data\safety\ held “the dry-run journal and stop file”. The STOP file is in data\security\. Anyone following the old line created a stop file where nothing reads one, and the kill switch did nothing at all — the app looks only at data\security\STOP. The exact path is shown in the app under Service → Emergency stop; that is the one to trust.

The log is the exception. It is written beside the executable, in the folder the installer put Autonoma in:

%LOCALAPPDATA%\Autonoma\bot.log

That is the file to send if you are reporting a crash. This page used to list a logs\ folder under .autonoma\ alongside the rest; nothing has ever written there, so anyone following it found an empty path and had nothing to send.

The encryption key for credentials is generated once, on your machine, and stored alongside them. No other installation can read your credential file.

What it will not do

  • It does not upload your content anywhere except to the platforms you connect.
  • It does not publish anything without a named person approving it first.
  • It does not spend money, place an order, or sign a transaction while dry run is on, or while the relevant feature switch is off.
  • It does not ship with anyone’s API keys, and it never sends your keys to us.

Something here wrong or missing? Tell us — documentation that describes software that no longer behaves that way is worse than none.