Run a company of one. Staffed like a hundred.
AI workers for sales, ops, research and content that run on your own machine. Agents prepare, you approve.
Your documents. Your AI. Nothing leaves the building.
Ask questions about your company documents and get answers with citations, from local models on your own PC. No cloud, no token fees.
A CMS and backend on Astro, in one Node process.
Content, a built-in admin, a REST API and an MCP server, so AI agents can read and write your content. GPL-3.0.
Rational text in irrational geometry.
Poems hidden in a double helix of coloured rungs, and a free offline phone app that reads them back, letter for letter.
Learn, build and grow with AI – together.
Courses, AI lessons, apps and live calls, with people building the same things. Opening soon at academy.theodorosdimitriou.com.
My thoughts, tutorials, and insights on web development, AI, technology & science.
AMD's tiny Ryzen AI Halo runs huge AI models without the cloud. It costs $3,999 with 128GB, and from $7,299 for the new 192GB model. The prices, the specs, the video of Lisa Su's demo, and an honest look at what fitting in memory does and doesn't mean.