Trajectory

Trajectory is the version control for your life: compare who you were with who you are becoming, trace the choices that shaped the change, and retain control over the story.

Comment

Inspiration

Life can have competing priorities - work, health, relationships, and personal growth - and while most trackers are good at telling me whether I completed a habit, they are less helpful when the question is larger: what is changing across my life, what led here, and what deserves attention next?

I built Trajectory as a private place to notice that longer arc. Trajectory connects daily reflection to a longer sense of direction.

What it does

Trajectory helps people see what is changing, decide what matters now, and choose where to go next - without streaks, guilt, or false certainty. It combines quick journaling, voice check-ins, weekly and monthly reviews, decisions, experiments, and a visual life journey

How I built it

I built Trajectory through an agentic collaboration with OpenAI Codex and GPT-5.6. Codex inspected the codebase, implemented full-stack features, audited authentication and privacy boundaries, debugged production issues, refined the responsive experience, and verified changes with tests. The stack includes Next.js, TypeScript, Neon PostgreSQL, Clerk, Drizzle, Gemini, and Vercel.

Challenges I ran into

The hardest challenge was turning scattered personal data into useful context without creating another overwhelming productivity dashboard. I also migrated from a local prototype to a secure multi-user application, protected sensitive journal data, and designed AI suggestions that assist without taking control.

Accomplishments that I'm proud of

Created a complete, deployed product with explicit user-scoped data access, guided onboarding, voice journaling, conservative AI-assisted capture filing, responsive design, backup and deletion controls, 46 unit tests, and 23 authenticated browser journeys. All within 3-4 days.

What I learned

Reflection needs context more than metrics. AI is most useful here as a quiet collaborator: reducing friction and organizing evidence without pretending to know what someone’s life means. Agentic development worked best as a continuous loop of inspection, implementation, testing, and refinement - not as one-shot code generation.

What's next for Trajectory

Next, I want to introduce private, user-controlled agents that can gather relevant journal entries, decisions, experiments, and life events into a draft monthly chapter; surface unresolved themes across competing priorities; and suggest small experiments aligned with the person’s chosen direction. These agents will remain transparent and approval-based: they can organize evidence and propose next steps, but never interpret a life, modify records, or act without permission.

Built With

Share this project:

Updates