Effective date: 19 October 2026
This policy is part of the Terms of Service. It applies to you, your members, your agents and your end users. If an agent breaks it, that counts as you breaking it.
1. Breaking the law
Do not use the Service:
- for anything illegal, or to help someone else do something illegal;
- in breach of sanctions or export controls. That includes use from, or for anyone in, a sanctioned country or region, and use for or by a sanctioned person;
- to infringe anyone's intellectual property, privacy or publicity rights.
2. People data and consent
- Look people up only when you have a lawful basis to do so.
- Hiring and court-record checks: run them only with the person's valid, informed consent, and pass that consent's record id. Never invent an id, reuse one, or pass one for a different person or purpose. You must first accept the Background Checks Addendum, and never use a check as the only basis for a decision.
- Do not use results to make decisions about credit, insurance, housing or similar eligibility.
- Do not use any superpower to:
- stalk, harass, threaten, dox or track a person;
- find a person who does not want to be found;
- build profiles on sensitive traits (health, religion, ethnicity, sexual orientation, politics, union membership, biometrics, precise location);
- discriminate unlawfully.
- Identity and fraud checks are for preventing fraud and meeting legal duties, not for surveillance.
- Do not resell or republish raw records as a data set, separate from your own product.
- Honour opt-outs and deletion requests for data you keep.
3. Messaging, SMS, email and calls
Follow Messaging and SMS in full. In short:
- message only people who agreed to hear from you;
- say who you are;
- honour STOP and unsubscribe at once;
- no spam, phishing or misleading sender names;
- no prohibited content (see Messaging and SMS).
Do not try to get around Pramaan's opt-out lists, send caps or kill switches.
4. Scraping and third-party services
- Respect each source's and each connected app's terms, rate limits and anti-abuse rules. When a superpower fetches public web data, use what you get lawfully.
- Do not use Pramaan to hide where automated traffic comes from, or to get around a site's access controls, paywalls or CAPTCHAs.
- Use only your own accounts and credentials as own keys, or ones you are allowed to use.
5. Fraud and abuse
Do not:
- open several accounts to collect starter credit more than once;
- pay with stolen cards, or file chargebacks in bad faith;
- impersonate any person, company or government body;
- use numbers, inboxes or domains for phishing, scams or fake reviews;
- launder money or move value through the Service.
6. The Service itself
Do not:
- probe, scan, overload or attack the Service, or try to reach another workspace's data;
- reverse engineer the Service, or try to find out which provider served a call in order to get around Pramaan;
- get around rate limits, spending caps, approval rules or key scopes;
- share keys outside your workspace (you may resell access to the API, including as your own API, but you stay responsible for every call made with your keys and for your customers following these rules);
- send malware, or content meant to make our systems or agents act against their instructions (prompt injection) aimed at other users.
7. AI use
Do not use model superpowers to make:
- child sexual abuse material;
- non-consensual intimate images;
- content that incites violence;
- deceptive deepfakes of real people;
- election disinformation;
- malware.
You must also follow the model providers' own usage rules: Anthropic and Google.
8. What we do about breaches
We may:
- refuse a call;
- pause a sending channel;
- revoke a key;
- suspend a member or the workspace;
- report illegal activity to authorities.
We will act in proportion to the harm and, where we can, tell you first.
9. Reporting abuse
Email abuse@pramaan.dev with what happened, the number or address involved, and when it happened.