Skip to content
Codezilla

Acceptable Use Policy

Last updated: May 2026

Codezilla v2 — Last updated: 26 May 2026

1. What this policy is

Codezilla v2 lets you run an AI agent that talks to your customers on WhatsApp and Instagram. That's a lot of power, and it sits on top of Meta's network — which means if you misuse it, you don't just lose your Codezilla account, you can lose your WhatsApp Business number permanently. This policy is the short, honest list of what's allowed, what's not, and what we do when somebody crosses the line. Read it once. It's written so a small business owner can understand it without a lawyer.

By using Codezilla, you (the tenant — the business that signed up) agree to follow this AUP. You also agree that you're responsible for every message your agent sends, even though the AI is the one typing it.

2. You ARE allowed to

Use Codezilla to:

  • Take orders — your customer asks for delivery, your bot takes the order, confirms the address, and sends it to your team.
  • Answer customer service questions — opening hours, prices, return policy, "is the chicken halal", "do you have size M", whatever your customers normally ask.
  • Book appointments — clinics, salons, tutors, mechanics. The bot reads your calendar and gives the customer a slot.
  • Answer FAQs from your knowledge base — upload your menu, your service list, your school's admission rules, and the agent will quote from it.
  • Send marketing to customers who opted in — anyone who clicked a checkbox, sent you a keyword like "JOIN", or otherwise gave you clear permission. Opt-in must be obvious, not hidden in fine print.
  • Re-engage past customers within Meta's 24-hour window — if a customer messaged you in the last 24 hours, you can message them back freely. Outside the window, you need an approved template and prior consent.
  • Coordinate internally with staff — using the agent as a back-office tool to assign tickets, log conversations, or notify your team.
  • A/B test your own messaging — within the limits below.

If you're not sure whether a use case is allowed, email [email protected] before you send. We will answer fast.

3. You are NOT allowed to

3a. Spam and unsolicited messaging

  • No sending bulk messages to people who didn't opt in. We don't care if you "bought a list" — buying a list is not opt-in.
  • No scraping phone numbers from Facebook groups, Instagram comments, OLX, or anywhere else.
  • No rapid-fire blasts designed to dodge Meta's rate limits. We monitor for this and we will throttle you.
  • No "growth hack" loops where the bot asks the customer to forward the message to 10 friends.

3b. Illegal goods, services, and content

  • No drugs, no prescription medication without a verified pharmacy license, no recreational substances.
  • No weapons, ammunition, or anything classified as a military item under Egyptian law.
  • No counterfeit goods. No fake brand-name products. No "first copy" Rolex.
  • No stolen goods, no smuggled goods, no goods that violate Egyptian customs law.
  • No gambling or betting platforms — including fantasy sports for money.
  • No adult content, escort services, or anything that violates Egyptian public morality law.
  • No content sexualizing or endangering minors, full stop. We report this to authorities immediately.

3c. Abuse, harassment, hate

  • No hate speech targeting religion, sect, nationality, ethnicity, gender, or any protected group.
  • No harassment of individuals — including ex-customers, ex-employees, or competitors.
  • No threats, intimidation, or extortion.
  • No doxxing — publishing somebody's private contact info or address through your bot.

3d. Deceit and impersonation

  • No pretending to be a bank, the government, Vodafone, Orange, Etisalat, WE, Fawry, Instapay, the tax authority, or any other entity you don't own.
  • No phishing — asking customers for OTPs, card numbers, national ID, or login credentials under false pretenses.
  • No fake giveaways or fake "you won" announcements.
  • No clone-shops pretending to be a real brand you're not authorized to represent.
  • No deepfakes of real people without their written consent.

3e. White-label, reselling, and abuse of the platform itself

  • No reselling Codezilla to other businesses as if it's your product. You can resell your services that happen to use Codezilla, but you can't pretend the platform is yours.
  • No white-labeling without a separate written agreement with us. Email [email protected].
  • No reverse-engineering, scraping the dashboard, or trying to break our rate limits, billing, or auth.
  • No sharing your login with people outside your business. Use the team-member seats.
  • No running multiple accounts to dodge a suspension. We will detect this.

3f. Egypt-specific prohibitions

  • No political campaign messaging unless you have the proper license from the relevant Egyptian authority. We will ask for proof.
  • No religious incitement, sectarian content, or anything that promotes hatred between Egyptians.
  • No content that violates Egyptian press and publication law, anti-cybercrime law (Law 175 of 2018), or telecommunications regulations.
  • No messaging that defames public officials, judges, or institutions in ways that violate Egyptian defamation law.
  • No content that promotes leaving the country illegally or smuggling people.

4. AI agent disclosure rules

Your bot does not need to scream "I AM A ROBOT" at the start of every conversation. But:

  • If a customer asks "are you a human" or "are you a bot" or anything similar, the agent must answer honestly. We've built this into the system prompt by default. Do not override it.
  • If a customer is making a serious decision based on what the agent says — medical advice, legal advice, large financial commitment — the agent should make clear it's an AI and recommend talking to a human.
  • You may give your bot a human-sounding name (Sara, Ahmed, Layla). That's fine. But the name doesn't get to lie about being a person.
  • You may not pretend the bot is a specific real employee of your company without that employee's written consent.

If you build a custom flow that violates the disclosure rule, that's a strike.

5. What happens if you break the rules

We use a three-strike system. A "strike" is any confirmed violation — confirmed by our abuse detection, by a Meta complaint, by an end-customer report, or by a manual review.

  • Strike 1 — Warning. Email to the account owner explaining what was detected and how to fix it. No service interruption. You get 7 days to respond and adjust.
  • Strike 2 — 24-hour cooldown. Your agent is paused for 24 hours. You can still log into the dashboard. We will tell you exactly what triggered the second strike.
  • Strike 3 — Suspension. Your account is suspended. We will cooperate with any Meta review and may share evidence with Meta as required by our Business Solution Provider agreement.

Some violations are one-strike-and-out, no warnings:

  • Child sexual abuse material
  • Phishing (impersonating a bank, government, telecom)
  • Coordinated political messaging without license
  • Reselling/white-labeling without agreement
  • Reverse-engineering the platform

Meta consequences are separate from ours. If Meta bans your WhatsApp Business number, we cannot get it back for you. That's why this policy is strict — we're trying to protect your number too.

6. How we monitor compliance

We don't read every message. We do:

  • Rate limits — if your agent sends more than the configured threshold per minute/hour, we throttle automatically.
  • ML-based abuse detection — we run lightweight classifiers on outbound messages for spam patterns, scam language, and prohibited keywords. False positives are reviewed by a human within 24 hours.
  • End-customer complaint reports — Meta forwards us reports when customers block or report your number. Repeat reports = strike.
  • Meta strike notices — Meta sends us quality-rating downgrades and policy violations. We pass those straight to your dashboard.
  • Random sampling — we audit a small random sample of conversations weekly for QA. We do not use your message content for any other purpose.

We log abuse signals in your dashboard under Compliance → Strikes so you can see exactly what's happening with your account.

7. Reporting violations to us

If you see another Codezilla tenant abusing the platform — spamming your number, impersonating you, scamming people — please tell us. Email [email protected] with:

  • The phone number or Instagram handle of the abuser
  • A screenshot of the message
  • The date and time

We will investigate within 48 hours and update you on the outcome (within the limits of what we can share about another tenant's account).

8. Tenant responsibility and indemnification

You are the publisher of every message your agent sends. The AI is a tool you operate. If a customer sues you, the government investigates you, or Meta penalizes you because of what your agent said, that is your responsibility, not ours.

You agree to indemnify Codezilla — meaning, hold us harmless and cover our legal costs — for any claim, lawsuit, fine, or penalty that comes from content you sent through the platform. This includes:

  • Claims by end-customers about scams, false advertising, or harmful advice.
  • Government fines for violating Egyptian law.
  • Meta's recovery of damages from us for violations on your account.

This indemnification does not apply if the violation was caused by a bug in Codezilla itself that we acknowledge in writing.

9. Changes to this policy

We may update this AUP as Egyptian law changes, as Meta updates its WhatsApp Business policy, or as new abuse patterns emerge. When we do:

  • We will email all tenant owners at least 30 days before the change takes effect.
  • The current version will always be at codezillaeg.com/legal/aup.
  • Continued use after the effective date means you accept the new version.

For urgent safety issues (e.g., new prohibition to comply with Meta or Egyptian law), we may make the change immediately and notify after, but we will still explain why.

10. Contact

  • General legal questions: [email protected]
  • Reporting abuse by other tenants: [email protected]
  • Appeal a strike on your account: reply to the strike email or open a ticket from your dashboard.

We aim to respond within 2 business days, faster for active suspensions.