Argus for journalists

Isolation

Open the page without introducing yourself

Every Argus profile starts from its own directory — separate storage, cache and cookie jar, with nothing carried over from any other profile. The session you open to read a company's careers page, or a forum you are watching for a story, knows nothing about the session where you are signed into your email, your CMS, or your own social accounts. Session restore is disabled, so a previous run cannot resurface in a new one.

That cuts both ways, and both matter: the subject of the story cannot connect the visit to you, and your accumulated browsing cannot contaminate what the site decides to show the visitor. You see the page a stranger sees, not the page your history has earned.

Fingerprints

The machine they see is not the machine you own

Sites do not need a login to recognize you — the hardware is enough. Canvas, WebGL, screen, timezone, CPU cores: together they name your laptop nearly as reliably as a cookie. An Argus profile presents a generated hardware identity instead — platform, memory, screen, timezone and languages produced together as one coherent set, so nothing contradicts anything else. The spoofing is applied in the renderer, below the page: a script sees the spoofed value because that is the only value present, not because a wrapper hid the real one.

Each profile keeps its identity across launches. The account you use to watch a Telegram channel looks like the same believable machine every morning — just never like yours.

Proxies · WebRTC

Read the story from where it is happening

Prices shown only to one country, a politician's ads targeted at one region, a page that serves different text to different geographies — verifying any of it means arriving from there. Assign the profile a proxy from the shared library and the whole session arrives from that geography. Argus health-checks every proxy and reports the egress IP, its country, and the latency — and a proxy that fails its check blocks the launch entirely, instead of quietly falling back and stamping your newsroom's IP onto the visit.

WebRTC is sealed as browser policy on every profile — a page cannot read your real address around the proxy, and it is not a toggle you can forget to set. Authenticated SOCKS5, the variant plain Chromium cannot speak at all, is handled inside the browser.

Profiles · Cookie sets

A persona that survives longer than one afternoon

Monitoring a community for months means the account doing the watching has to stay consistent: same identity, same sessions, same logins. In Argus that persona is a profile — its cookies live in the profile's own jar, and a login saved as a cookie set is seeded into the session at launch, so the profile opens already signed in every time.

A beat becomes a folder of profiles: each row with its own status label and notes — which source it watches, when it was last used, what it must never be used for. Deleting is soft, so a persona retired at the end of a story can be restored if the story reopens.

Cloud sync · Teams

The investigation is not trapped on one laptop

Sign in with an email and password and your workspace follows you — the profiles, proxies and cookie sets are synced to your organization, not stranded on the machine where they were made. A desk working one investigation shares one pool: the colleague picking up your monitoring list opens the same rows you left, and two people working at once cannot overwrite each other.

When a story hands over — a reporter rotates off, a partner outlet joins — the handover is an account, not a folder of exported cookie files traded over email.

Reach · Datasets

Build the source list first, then work it under a persona

Before you can watch a beat you need the list: the channels, the subreddits, the feeds, the accounts worth checking every morning. Gathering that does not need a browser session, and Argus has a step that does it without one. Reach pulls a page through its own channels — a YouTube channel, a subreddit or an RSS feed answer as structured rows, one per video, post or item, and an ordinary page answers as readable text.

It runs from the launcher's own connection, which is the trade worth understanding: it is fast and costs you no session, but it does not use the profile's proxy or cookies, so the site sees your real address. Use it for public feeds and pages you would not mind being seen reading — and a browser profile for anything else. A login-gated page needs the browser regardless.

Save what it finds into a dataset and the list stops being a scratch file on one laptop: a table your workspace owns, with columns you named, that you can sort, filter, tag and hand to the colleague covering for you. Put the gathering step on the calendar and it refreshes itself before you are at your desk.

Then work the list under a persona. The rows that need an account behind them — the community that hides its posts from logged-out visitors, the group you had to be admitted to — get opened by the profile you built for that beat, with the same fingerprint, the same cookies and the same proxy every time. The two halves are deliberately different tools: one is cheap and anonymous, the other is careful and consistent, and spending a persona on a task that never needed one is how personas get burnt.