Privacy and Security

Argus is built to keep your data on your own machine. This page lays out what it reads, what it stores and the few times anything leaves your computer.

Open source and local

Argus is a free and open source tool, MIT licensed, so you can read exactly what it does. It runs entirely on your own computer. Installing it and browsing your usage needs no account and no server.

Indexing and storage

Argus finds and indexes the sessions your AI agents have already saved to your own disk (see Supported Agents for the locations it reads). Indexing is entirely local: it reads those files, pulls out the useful details and saves them to a local index, a database on your computer. It doesn't watch your screen or record anything as you work. Everything it needs is already on disk.

Credential warnings

As it indexes, Argus checks your session text for credentials you may have pasted into a conversation, like an API key, a token or a private key. The check is pattern matching on your own machine: no model reads it, and nothing is sent anywhere.

What Argus keeps from a match is deliberately thin. It records the kind of credential, where in the session it appeared and a few characters of the value, enough to tell you which key it was and not enough to use it. The credential itself is never copied into the index.

Warnings stay on your computer. A sync uploads none of them. What an Argus Hub does receive is one yes or no on each task, saying whether a warning landed in it, so whoever runs the Hub can see that a piece of work touched a credential and ask you to rotate it. The kind, the location and those few characters all stay local. Dismissing a warning in your own Argus hides your banner and leaves that flag in place, since the flag records that the work touched a credential rather than that you've read the warning.

Task interpretation and model providers

Task interpretation is the one part of Argus that uses a model to read your sessions. It's on by default, and you choose which model provider does the reading, or turn it off entirely. To judge how a task went, Argus sends that task's prompts and responses to the provider you chose, so some of your session content is shared with that provider. The default sends it to Anthropic through your Claude sign-in; the hosted providers send it to their own services. Only a local model run through the Command provider keeps it on your machine. See Tasks for the full picture.

Agent access

Argus can let AI agents on your computer query your data themselves over a local MCP endpoint, so an agent can answer questions about your past work. It's on by default, and it's read-only: an agent can look at your sessions, usage and task outcomes, but can't change anything. The endpoint only answers on your own computer (it refuses requests that don't come from localhost), and Argus still sends nothing anywhere itself.

The one sensitive piece is transcripts. Reading one sends the full text of that conversation into the agent's model provider's context, so transcript access ships off. Turn it on under Settings → Agent access only if you're comfortable with that. See Connect Your Agent for how it works.

API keys

When you use a provider that needs an API key, Argus stores the key in your operating system's secure store (the Keychain on macOS), never in its settings file and never on any server. It reads the key from there only to call the provider you configured.

Updates

The desktop app checks for new versions on a schedule and downloads them in the background. The Updates setting in Settings controls whether new versions automatically install; with it off, Argus still checks and tells you when one is ready, and you install it from the menu bar. The command-line tool doesn't check for updates at all.

Those checks and downloads go through a download service The Agent Deployment Company runs, and they carry one detail about your copy of Argus: a random identifier it creates on your computer the first time it runs. It lets us count how many installs are checking for updates, rather than counting the same computer over and over. It isn't tied to your name or your email, and nothing about your sessions, your settings or your usage goes with it. New versions themselves are hosted on GitHub, so when Argus downloads one, that identifier reaches GitHub along with the request. If you sync to an Argus Hub, it's the same identifier the Hub uses to tell your uploads apart from a colleague's.

If you'd rather not be counted, turn off Share update metrics in Settings. Argus still checks for updates and still installs them; the checks just don't say which install they came from.

Argus Hub

By default, Argus uploads nothing. If your organization runs an Argus Hub, you can opt in to sync a snapshot of your usage to it, so an ops leader can see agent use across the team. Two things to know:

• The Hub is hosted by your own organization, not by The Agent Deployment Company. Your data goes to your company's server, not to us.
• What's sent stays minimal. A sync carries your metrics and task data (usage totals, breakdowns, outcomes and the like), plus a few short text snippets: a session's opening prompt and the brief evidence behind a task's judgment. The full text of your sessions is never uploaded, and neither are the details of a credential warning, only a flag on the tasks one applies to.