penn AI

A human-in-the-loop X reply copilot that drafts replies and posts in your voice, using your context only when relevant.

penn AI is a human-in-the-loop reply copilot for X. It lives beside the X composer. When you click Suggest replies, it reads the post you are replying to, combines it with your saved profile (who you are, what you build, your actual posts, your forbidden phrases), and drafts 3-5 reply options that read like a sharp human typed them on a phone. WHY IT DOESN'T SOUND LIKE AI Every draft runs through a filter that bans the tells people clock instantly: the "not X, it's Y" contrast flip, em dashes, rule-of-three lists, hype words, listicle voice, engagement bait, forced wrap-ups. If a draft sounds like a model wrote it, it never reaches you. YOUR PRODUCTS, MENTIONED ONLY WHEN IT FITS Save your products and projects with rules for when to mention them. A relevance gate decides honestly whether a mention belongs in the reply; a forced plug is worse than no mention, so most of the time the answer is no. GROW WITH PRO - Free, forever: 5 generations a day, replies in your voice, anti-AI-slop filter, reads images in the post. - Pro ($9/mo or $79/yr): 400 generations a day on the best model, original posts grounded in your live feed and trends, builder-voice product promotion, web search for current events, draft refinement chat, model choice. PRIVATE BY DESIGN Your profile stays in your browser's local extension storage; our server never keeps a copy. Generation requests are processed and discarded — request content is never logged or stored. Sign-in is Google; payments are handled by Polar. No ads, no trackers, no data sales. HARD SAFETY LINES penn AI never auto-posts, never likes, follows, DMs, or reposts, never scrapes, and never acts without your click. You always edit and post by hand.