Build

Policy — Platform Restricted Terms (IG/TikTok Safe-Words)

3 min read

Policy — Platform Restricted Terms

Purpose: keep Indexa’s Instagram and TikTok accounts alive. Certain words get content suppressed (shadowban) or get the whole account banned — especially named pharma drugs, injection/dosing language, and medical claims. This policy defines what we never say on-platform, what we say carefully, and how the automation enforces it so a bad post can’t go out by accident.

This is enforced in code. The publisher (Automation/Zernio/publish.mjs) runs a lint gate before uploading anything. A block term aborts the whole run. The word lists live in Automation/Zernio/restricted-terms.json — edit that file to change the rules (single source of truth). The standalone checker is node lint.mjs.

Sits under the existing guardrail in SOP — Zernio Publishing and Distribution — Strategy (research-use framing, no medical claims). This policy is the platform-survival layer on top of the legal layer.

Two tiers

BLOCK — never on-platform (automation refuses to publish).

  • Named pharma drugs: Retatrutide, Tirzepatide, Semaglutide, Liraglutide, Cagrilintide, Ozempic, Wegovy, Mounjaro, Zepbound, Saxenda. Use a code name or “the compound” instead.
  • Administration / dosing: inject, injection, needle, syringe, subcutaneous, intramuscular, dosage, dose, dosing, mcg, reconstitute, reconstitution.
  • Medical claims: cure, cures, treats, treatment for, heal disease, prevents disease.
  • Consumption framing: “for human use”, “human consumption”, “safe to take”, “take this”.

CAUTION — allowed, but flagged every time (use sparingly, swap in hashtags where you can).

  • peptide(s), vial(s), research chemical(s), weight loss, fat loss, muscle growth, anti-aging, SARMs, steroid, hormone, clinical, mg, ml, protocol, stack.

“Peptide” is unavoidable in our niche — it stays caution, not block. The play is to reduce its density in captions and hashtags (the parts platforms scan hardest), not to purge it from educational copy. On-image text is lower-risk but still counts.

Approved swaps (say this instead)

Don’t saySay
Retatrutide / Semaglutide / TirzepatideReta / a code name / “the compound”
peptide(s)research compound(s) / “the compound”
weight loss / fat lossmetabolic research / body-composition research
injection / injectresearch handling / preparation (route detail off-platform)
dosage / doseresearch protocol amount
reconstitutionpreparation steps (link off-platform)
buy / for saleenquire / DM to order

Where this applies

Captions, hashtags, first comments, bio, and on-image text (platforms OCR images/video frames). The linter scans caption, onImageText, and note for every post in posts.json — so keep onImageText in the post entry accurate to what the creative actually shows.

How to use it

  • Before scheduling: node lint.mjs (or it runs automatically inside node publish.mjs). Fix every BLOCK; review every CAUTION.
  • Override (rare): node publish.mjs --force publishes despite block findings — only for a deliberate, reviewed exception.
  • Tuning: add/remove terms in restricted-terms.json. Promote a caution word to block if a platform starts penalising it; demote if proven safe.

Operating rules

  1. Lead generation happens off-platform. On IG/TikTok we educate and build trust; product names, prices, and ordering move to WhatsApp/link-in-bio. This keeps the riskiest words out of public posts entirely.
  2. Never put a price or “buy now” in a caption — route to DM.
  3. Keep the research-use disclaimer on every post (legal layer) and keep block words out (platform layer). Both, always.
  4. Threads uses disposable alias accounts (separate playbook) — assume even stricter and rotate.

SOP — Zernio Publishing · Workflow — Content Production Pipeline · Style Guide — TikTok Braindead Poster · Style Guide — Instagram Beautiful · Automation/Zernio/restricted-terms.json

Graph View

Backlinks