# Safety & compliance

What separates a legitimate outreach tool from a spam bot is its defaults. This is what Adnetic does out of the box — nothing to switch on.

Updated: 2026-10-13 · https://adnetic.org/en/safety/

## Six non-negotiable rules

1. **Public data only** — No logging in to read private content. For forums, Adnetic reads each site's robots.txt and terms, and refuses to crawl where the terms forbid it — even when robots.txt allows it.
2. **A human approves before posting** — Every reply goes to an approval queue. Auto-approve is off by default; posting adapters run in DRY_RUN until you deliberately switch them on.
3. **Ads are labelled** — Replies automatically carry a public disclosure — #quảng_cáo by default, or a label such as (sponsored).
4. **Rate limits** — Daily caps per campaign, rate limits per account and platform, with backoff when a platform pushes back.
5. **No repeat targeting** — Deduplication by author and post, with a cooldown before anyone can be contacted again.
6. **Global opt-out** — People who appear in the data can ask not to be contacted. The request applies to every Adnetic customer and is filtered before classification even runs.

## Posting through official APIs only

Adnetic never acts on your behalf through a platform's UI: no browser-automated posting, no auto-joining groups, no following. Posting only goes through official APIs with the account owner's consent. See [supported platforms](https://adnetic.org/en/#platforms) for where posting is possible.

## Controls & audit log

- A brand-safety keyword blocklist: matching posts are flagged and never become a lead for any customer.
- An activity log: who created a campaign, who approved or rejected a reply, when the plan changed.
- Sign-in through Keycloak (OIDC), with advertiser-admin, reviewer and viewer roles.

## Requesting an opt-out

If your post or comment shows up in Adnetic and you don't want to be contacted, send an opt-out request with the platform and your account ID. No login is needed, and the request applies to every customer. Technically: `POST /opt-out` on the Listener API with `platform`, `author_id` and an optional `reason`.

## Safety questions

### Can using Adnetic get my account banned?

Posting through official APIs, with the right permissions and rate limits, is the lowest-risk path. Browser-based reading breaks Meta, TikTok and Threads terms, so it is off by default — turning it on is the product owner's call.

### Is Adnetic a substitute for legal advice?

No. Adnetic ships safe defaults, but you still need to review advertising and personal-data law in your own market.