IntentIQ
Record Once. Train Never. Test Anywhere.
Teaching IntentIQ
Blueprint Engine
Core Idea
Playwright
Execution
BrowserUse
Self-Healing
Project overview
I wanted to build a smart bot. It needed to work everywhere, heal itself, and require minimal human intervention.
Initial Thought
I started with a simple goal: build a smart, self-healing bot that works across every web application without writing scripts for each one. It needed to minimize API calls and require minimal human intervention.
Turning Point
Then I realized: Humans don't think in selectors. Humans think in workflows. That realization became the turning point. I didn't want one file `smart-bot.py`. I needed a system.
Architecture
OpenAI became the Brain. Playwright became the Hands. BrowserUse became the Doctor. Recorder became the Teacher. Blueprints became Memory. Judge Engine became the Critic. Learning Engine became the Student.
Things I Didn't Expect to Build
Outcome
I started with a simple goal: Build a smart bot. Somewhere along the way, I accidentally built: • A Recorder. • A Blueprint Engine. • A Judge Engine. • A Learning Engine. • A Competition Engine. • A Workflow Repair System. • Native OS-level popup handlers. IntentIQ taught me an important lesson: > Software is easy. > > Teaching software to survive in constantly changing environments is hard. And if I had to summarize IntentIQ in one sentence: > "I didn't build a QA tool. I built an AI agent that learns how applications work."