Who this applies to
Everyone who uses the platform, and everything built on it: agents, workflows, widgets, connections and the content they produce. If an agent does something on your behalf, we treat it as if you did it, so set up its tools and approvals with that in mind.
Never
Do not use MyDeepAgents to create, send or help with:
- anything illegal where you or the people affected are, or that breaks someone's rights, including privacy and intellectual property;
- child sexual abuse material, or anything that sexualizes minors, which we report to the authorities;
- threats, harassment, hate, or encouragement of violence or self-harm;
- scams, phishing, fraud or impersonation of a real person or organization;
- spam, including bulk messages to people who did not ask for them;
- malware, or tools to break into, disrupt or take over systems;
- personal data you have no right to collect, or tracking people without a lawful reason;
- election interference or deceptive political messaging;
- content the AI model providers' usage policies forbid.
Be clear that it is AI
When an agent talks to people, through a widget, email, chat or any other channel, they must be able to tell it is an AI and not a person. Do not tell an agent to claim otherwise. If someone asks, the answer must be honest.
Decisions with serious effects
Do not let an agent alone decide something that seriously affects a person, such as credit, insurance, employment, housing, education, legal status, or medical diagnosis and treatment. A qualified person must review those decisions. Approvals on the platform exist for this: use them.
Respect the platform
- Do not probe, scan or test the security of MyDeepAgents except as our security page allows.
- Do not get around usage limits, access controls, billing, or the separation between organizations.
- Do not overload the service, or use it to overload anyone else's.
- Do not resell or copy the service to build a competing one.
Connected tools and models
When agents act in a connected tool or call an AI model, follow that company's own rules too. Connect only accounts you are allowed to use, and give agents no more access than the work needs.
Report misuse
If you see MyDeepAgents being misused, including by an agent or widget on another website, tell us at khurram@mydeepagents.com. Include links or screenshots if you can. We look into every report.
What we do about it
Depending on how serious it is, we may warn you, remove content, pause an agent, widget or connection, or suspend or close the account. Where it is fair to, we will tell you first and give you a chance to fix it. Where the law requires, we report to the authorities.
Changes to this policy
We may update this policy as the platform and the law change. We will update the date at the top, and tell you about important changes before they take effect.