Acceptable Use Policy

1. Responsibility for automation

You are responsible for every message and telecom action made through your workspace, whether a person, script, integration, model, or agent made it. Give software the narrowest permissions it needs. Add human approval where a wrong recipient, repeated message, or misunderstood reply could cause harm.

2. Messages you may send

You may send conversations started or expected by the recipient, service messages such as appointment updates or security codes, and marketing messages where you hold valid permission. Every marketing message must identify the organisation and give a simple way to opt out. Keep evidence of the permission and any later withdrawal.

3. Prohibited content

You must not use Agent Telco for messages involving:

• fraud, phishing, smishing, impersonation, or deception;
• malware, credential theft, or deceptive links;
• threats, harassment, hate speech, or unlawful discrimination;
• unlawful goods, services, or content;
• unregulated loans, debt relief, cryptocurrency schemes, or other high-risk financial promotions;
• gambling, tobacco or vaping products, prescription-only medicines, adult services, weapons, or illegal drugs;
• third-party debt collection or credit repair services.

4. Prohibited practices

• sending unsolicited bulk messages;
• using bought, rented, scraped, or unlawfully shared lists;
• continuing after an opt-out or disguising the sender;
• rotating numbers, workspaces, keys, clients, or endpoints to evade filters, limits, suspension, or provider review;
• probing, attacking, or disrupting Agent Telco, telecom networks, recipients, or third-party systems;
• placing API keys, OAuth tokens, webhook secrets, or personal data into public prompts, logs, source code, or model-training data;
• reselling the service without our written agreement.

5. Consent and opt-outs

Consent for text marketing must cover your organisation and text messages. A phone number being public, an inbound message, or generic permission to market is not enough by itself. If the law permits a soft opt-in, you remain responsible for meeting every condition and giving an opt-out when details are collected and in each message. Stop messaging promptly when a person withdraws permission.

6. Security and webhooks

Store credentials outside prompts and logs. Verify webhook signatures against the raw body, reject stale timestamps, use HTTPS endpoints, and deduplicate event IDs. Revoke a credential promptly if it may have leaked. Do not let an agent reveal credentials or approve its own increase in authority.

7. Network protection

Mobile networks apply their own filtering. We may throttle, reject, quarantine, or pause traffic when patterns, complaints, costs, or provider feedback indicate abuse or a compromised agent. Do not split traffic to work around those controls.

8. Enforcement

We may ask for consent records, agent controls, or an explanation of traffic. We may warn, restrict, suspend, or terminate access in proportion to the risk. Clear fraud, unlawful traffic, credential compromise, or immediate recipient harm may lead to suspension without notice. We may preserve evidence and report suspected crime or network abuse where the law permits or requires it.