How Publications Are Made: IPs, Accounts, Behavior
“Publications from different IPs” is a minimum, not an advantage: mere IP diversity is no longer sufficient. Below is what actually happens with each publication, including those layers that are usually not discussed: browser fingerprint, behavior, captchas, and registration email.
Publication goes through a real browser
The placement is not an HTTP request to a form, but an automated Chromium: the login page opens, fields are filled, the same site scripts are triggered as with a live user. This is noticeably more expensive than a simple POST request, but this is how publication passes where there is at least some protection, and it looks like a normal session, not a bot.
Automation is launched with an anti-detect layer: signs of a controlled browser are hidden, and the environment — window size, language, time zone, capabilities set — looks like an ordinary user configuration, not like a default headless.
Network layer
| Channel | When used |
|---|---|
| ISP proxies — static addresses of real providers | Main publication channel: stable address, live session, cookies are preserved |
| Residential rotating addresses | Registrations and multi-step scenarios where a fresh exit is needed |
| Antibot gateway | Retries and sites with strict protection: takes on the passing of checks |
Each proxy is checked for liveliness before use; non-working ones are marked and excluded. Publications from the server IP of the platform do not go through — such an address quickly becomes recognizable.
Captchas
Some sites are closed with checks: Cloudflare challenge, hCaptcha, reCAPTCHA, Turnstile. For them, proprietary solvers and external recognition services are used. Where the check cannot be passed consistently — the site is simply excluded from the pool, rather than remaining “sometimes working”: half of the failed publications are worse than the absence of the site.
Accounts and email
- Accounts are registered on a proprietary email domain with reception on all addresses — registration confirmation comes and is processed automatically.
- There is a limited number of publications per site within one launch: the reuse parameter for the site is set by settings and does not allow the entire cascade to go to one domain.
- Accounts and sessions are stored and reused: an account with a history of publications looks more natural than a newly created one.
Publication tempo
The cascade does not publish everything in one hour. In natural mode, publications are distributed unevenly over time, in normal mode — they go in parallel streams with a limit on simultaneous tasks. The average duration of a completed launch is about 17 hours (based on 286 launches since July 2026).
What happens after publication
- The publication address is saved in the report.
- A check is launched: the page is reopened and your link is searched for — see why only verified links are in the report.
- The publication is sent for indexing, the status is checked in search — see indexing of publications.
- If the slot is not confirmed, it is reallocated to another site at our expense.