Policies
Submission guidelines
How to submit a prompt to AIPrompt123: originality, formatting, safety review and what happens after you hit send.
Last updated
We accept original prompts from practitioners. Every submission is read by an editor, tested on at least two models where feasible, and either published with credit or returned with notes.
What we look for
- Original work that you wrote or have permission to share.
- A specific job, not a generic request such as 'write a blog post'.
- A clear output format, audience and set of variables.
- A worked example of the input and the kind of output it produces.
- Notes on failure modes: what the model gets wrong and how the prompt prevents it.
What we reject
- Prompts copied from another site, book or repository without permission.
- Content that helps with malware, fraud, harassment, self-harm, sexual content involving minors, or evading safety systems.
- Medical, legal or financial prompts that present model output as professional advice.
- Prompts that depend on a leaked system prompt or a jailbreak.
- Low-effort variants of an existing template with no meaningful change.
How to format a submission
Include the prompt text exactly as it should be copied, the variables in curly braces, the platforms you tested, the model versions, a short example input and an example output. If you have a screenshot, attach it; do not send personal data in examples.
Review timeline
We aim to respond within ten business days. Accepted prompts may be edited for clarity and formatting; we send substantive changes back to you before publication. If we decline, we explain why so you can revise and resubmit once.
Credit and licensing
Accepted prompts are published with your name and a short bio if you provide one. You keep ownership and grant us a non-exclusive licence to publish, edit and promote the prompt. We do not pay for individual submissions; the newsletter and premium programme may offer separate paid opportunities.