Split screen: raw JSON draft output on one side, polished LinkedIn post on the other, showing the editing process.

This post was drafted by an AI running on hardware in my garage. I think you should know that.

Not ChatGPT. Not Claude. Not any cloud API. My own agent, named Mr. Peepers, running a 30B parameter model on my own GPU, in my own house.

Here's how it works: Mr. Peepers has a sub-agent named Velma (yes, from Scooby-Doo). Velma handles content and marketing tasks. When it's time for a social media post, Mr. Peepers delegates the writing to Velma, who drafts the LinkedIn post, the Twitter version, suggests hashtags, and even describes what the accompanying image should look like.

The drafts land in a content management system I built — a Flask app running in Docker where I can review, edit, approve, or reject each piece. Nothing posts without my approval. The AI proposes, the human disposes.

Why build this instead of using ChatGPT? Because I can. Because the data never leaves my network. Because customizing the model's personality and knowledge base is trivial when you control the whole stack. And because it's genuinely fun.

The future of content isn't AI replacing humans — it's AI doing the first draft so humans can focus on the parts that matter.

#AI #ContentCreation #LocalAI #HomeLab #Automation #LinkedIn #OpenSource

Posted with love and responsible renewable AI accountability with Mr Peepers AI Bot