What we're building.

Public read-only endpoint so ChatGPT, Claude, and other agents can fetch a user's pillars + active goals to stay aligned with what the human is actually working on.

Mint a bearer token from settings, paste into your AI's system prompt, and the agent sees your full alignment context — drains, parked, projects, habits, current focus — not just the public shape.

Zero-dependency Node MCP server. Drop pathnav-mcp.mjs into Claude Desktop or Cursor, paste your token, and the agent gets get_alignment + get_alignment_markdown tools it can call on its own.

Authorized agents POST proposed plan changes via Bearer token; pending queue at /dashboard/proposals; approve auto-applies the change locally and syncs up via the snapshot. Five kinds auto-apply (milestone add/complete, drain add, parked add, set_focus); notes still record-only.

Sketch a 5-year vision in plain English; PathNav drafts pillars, goals, and milestones for you to edit.

Map your plan as a gantt, a calendar, a moodboard, or a freeform canvas. Switch between views without losing structure.

Every morning, a one-screen brief: news that affects your plan, blockers to clear, and the question to chew on today.

Practice tracking with streaks. Drains are explicit anti-goals so the Guide steers context away from things you're trying to drop.

Opt-in public profile with vision + active goals, custom OG card, anonymous cheers + notes from visitors.

Browse opt-in public profiles sorted by cheers. Real visions, real plans.

Each completed milestone gets a green-gradient celebration page with custom OG card. Share the win without exposing your whole plan.

Pro plan $19/mo or $190/yr. /upgrade route resolves session and routes you straight into the right checkout.

Add your own AI provider keys for unlimited use. Provider + model picker on every AI action.

Long-form planning units shared between users. Adopt a Tract, steward it for a season, hand it off cleanly.

Live readiness across storage, worker, AI, billing, email. 60s refresh. Incident timeline.

Token scopes (read-only · propose · auto-apply) gate trust level. Notes feed surfaced on the dashboard so add_note proposals close the loop. Webhook outgoing so agents can be pinged on plan changes.

Speak a quick note, hear your morning brief on the way to work. Web Speech API for capture + synthesis with intent matching.

Schedule by what you actually have capacity for, not just what's on the list. Workbench rows pick up the new fields automatically.

Follow people whose plans inspire you. Activity feed of milestones their visible profiles complete. Privacy-by-default — only opt-in profiles surface.

Households can share a subset of pillars + goals + practices. The Family plan unifies parent-managed and adult accounts under one calendar.

Opt-in daily summary with the one focused question for the day, weekly progress retro, and any cheers/notes you received.

Side-by-side monthly vs annual cards with a single toggle. Annual savings highlighted.

Free users see a tiny 'X of Y free AI calls used today' meter with a one-tap upgrade link.

Post-deploy smoke script that hits public endpoints + a few authenticated paths against a test account, exits non-zero on regression.

Daily encrypted snapshot of state to a second region so a disk-loss event doesn't take user data with it.

Browser-driven tests against signup, onboarding, plan creation, share flow. Catches regressions vitest can't.