After a decade advising mental health companies, I founded my own. Loop Mind is a voice journaling app that helps chronic overthinkers see the patterns in their own thinking — built solo, shipped to the App Store, and run with an AI-powered operating model. Live today at loopmind.care.
Project Snapshot
Loop Mind
Voice journaling app that detects thinking patterns
2026 – Present
Founder
Product strategy, AI system design, full-stack delivery
Solo founder, operating through an AI agent model
Live on the App Store since May 2026
Strategic Insight
Overthinking is one of the most common mental health complaints — replaying conversations, spiraling before decisions, worrying about symptoms. Yet almost every app in the space responds the same way: journaling prompts, mood tracking, or a chatbot playing therapist.
None of them answer the question overthinkers actually have: what is my mind doing, and does it keep doing the same thing?
Loop Mind is not a journal, not a therapist, and not a mood tracker — it's a thinking tracker. You talk out what's looping in your head, and the app shows you the pattern: what kind of loop it is, what triggers it, and what keeps coming back.
Why I Built It
At Mentalyc, I spent 18 months designing systems that turn therapy conversations into clinical signals — therapeutic alliance, symptom patterns, treatment progress. The work convinced me of something bigger:
- Spoken thought is the richest mental health data there is — far richer than a mood rating or a typed entry.
- Most of that thinking happens outside the therapy room — between sessions, or for people who were never in therapy at all.
- AI had just become good enough to reflect thinking back to a person honestly, without pretending to be a clinician.
Loop Mind is the direct continuation of that thesis, applied to the consumer side — and this time, I wanted to own the outcome rather than advise on it.
The Product
The core loop is deliberately simple: when your head is spinning, you talk. Speaking is lower friction than writing when you're spiraling — and it preserves the raw texture of the thought.
Voice-first capture
Press one button and talk. AI transcription handles the rest — no typing, no forms, no friction at the exact moment friction matters most.
Honest reflection
Each entry comes back as a structured reflection: the main loop it detected, what you may be feeling, and a fact-versus-story breakdown of the thought.
A clinically grounded taxonomy
Every entry is classified into one of seven research-backed loop types — rumination, anticipatory worry, somatic (health) anxiety, and more — so patterns become nameable.
Patterns over time
Recurring themes, common triggers, connected beliefs, and intensity trends build up into a picture of how your mind actually behaves.
Clinically Grounded, Never Clinical
The hardest design constraint was positioning. Loop Mind sits next to therapy, not in place of it — its stickiest use case is people in therapy using it between sessions, because "it shows them things."
That constraint became a product test applied to every screen and every sentence of copy:
✓ Loop Mind territory
"See what your brain is doing" — observation, patterns, honest mirroring.
✕ Not Loop Mind territory
"Feel better about what your brain is doing" — treatment, advice, therapy claims.
The loop taxonomy is grounded in clinical literature — the somatic loop, for example, follows the health anxiety model of Salkovskis & Warwick — and it only earned its place in the product once real usage data showed enough entries fit it. The science is in the structure; the voice stays human.
One App Store screenshot carries the whole positioning in four words: "A mirror, not a therapist."
Building Solo, With an AI Team
Loop Mind is also an experiment in how a product gets built. There is one human on this team. Everything else — engineering, QA, analytics, marketing, product management — runs through a structured system of AI agents I designed, each with its own mandate, memory, and rules.
Departments, not prompts
A chief-of-staff agent routes work to specialist agents: an engineer that follows architecture and security rules, a QA agent that clones fresh and tests before anything ships, a CTO for strategy, a CMO for positioning.
Governance that survives speed
Every analytics event must name the decision it informs before it can exist. Every PR passes a quality gate. Rules live in shared memory, so the standard holds even when shipping daily.
Infrastructure a team would envy
Over-the-air updates ship JS changes without an App Store cycle. Crash reporting, product analytics, subscriptions, and lifecycle email run as an integrated, audited stack.
1
Human on the team
7
Research-backed loop types
May 2026
Live on the App Store
What This Project Represents
The thesis, productized
Ten years of consulting on mental health technology, condensed into one product I fully own — from strategy to App Store.
AI as infrastructure, not therapist
A working demonstration of the positioning I advise clients on: AI that strengthens human insight instead of simulating care.
A new way of building
Proof that one person with a well-designed AI operating model can ship and run what used to take a team.
Loop Mind is live today. Explore it at loopmind.care or find it on the App Store.
Building a product intelligence function?
If you're building in mental health, digital health, or regulated care and don't have a product function yet, or have one that's not working, I'd love to hear about it.