Own the pipeline. Outsource nothing that matters.
Glad Labs builds the tooling and the playbooks for people who want to run their own AI publishing pipeline — without handing the content, the relationships, or the vendor-lock-in to a SaaS company.
The product is Poindexter: an open-source autonomous publishing agent that runs one post a day at a quality score above 80, in our house style, on one enthusiast-grade PC. Local by default. No one else deciding what you can write about.
Poindexter Pro is how you get there fastest — the operator console, the live-tuned configuration from the running system (rebuilt weekly), the full operator book, and the founding Discord. The Writing & Research site (gladlabs.io) is where we show our work.
Every step can run on hardware you own: Ollama, local image generation, local speech. Hosted models are opt-in behind a spend cap — our own instance rents one, the writer, for about $10 a month, and runs every other call locally.
Fork the entire pipeline. Run it. Modify it. Sell services or a SaaS around it. Permissive license with a patent grant — built to be integrated, not just admired.
Engine is free and open-source. Poindexter Pro is $19/mo or $180/yr (Founding Member rate, locked for life). Cancel anytime, keep everything downloaded. Stripe-less — checkout runs on Lemon Squeezy.
PostgreSQL. Next.js. FastAPI. Ollama. Prometheus + Grafana. Everything else is a liability.