"To give soul to your tasks" — A Smart AI Task Manager with transparent reasoning.
Jaanify is an AI-powered task manager that shows you why it makes each recommendation, not just what to do next.
Most AI task managers operate as black boxes — they reorder your priorities without explanation. Jaanify takes a different approach with 3-Tier Reasoning Cards: progressive disclosure that lets you glance at a recommendation, explore the factors behind it, or audit the full reasoning chain.
Live demo: jaanify.vercel.app
Built for: Neurodivergent individuals (especially ADHD), freelancers, and solopreneurs who need transparency and trust from their tools.
- Natural Language Task Input — Text and voice capture with real-time AI parsing
- 3-Tier Reasoning Cards — Glanceable, explorable, and auditable AI explanations
- AI Daily Planner — Personalized task ordering with visible factor analysis
- 60-Second Onboarding — Start with "What's on your mind?", account creation deferred
- Voice Capture — Tap-to-speak task entry in 8-12 seconds
- PWA-First — Web + offline-capable from day one
jaan-to is a Claude Code plugin that provides 21 AI-powered skills for product workflows — from writing PRDs and user stories to generating frontend designs, backend task breakdowns, API contracts, test cases, and more.
Skills span six domains:
| Domain | Examples |
|---|---|
| Product Management | PRD writing, user story generation |
| Development | Frontend/backend task breakdown, API contracts, data models, stack detection |
| UX & Content | Microcopy packs, research synthesis, heatmap analysis |
| QA | BDD/Gherkin test case generation |
| Data & Analytics | GTM dataLayer tracking |
| Workflow | Roadmap management, learning capture, skill creation |
This project serves a dual purpose:
- Build a real product — Jaanify is a launchable task manager, not a demo
- Stress-test jaan-to — Every artifact is produced exclusively through jaan-to skills
Jaanify and jaan-to evolve together in a structured cycle:
SCAN → REVIEW & TEST → BUILD → GAP REPORT → (user adds skills) → repeat
- Scan — Pull latest jaan-to, check for new or updated skills
- Review & Test — Run each skill against Jaanify context, score results on per-skill scorecards
- Build — Advance Jaanify using all available skills
- Gap Report — Document missing capabilities, request top priorities
Each cycle is tracked with version-pinned scan reports, scorecards, and gap analyses in this repo.
| Phase | Status | Key Outputs |
|---|---|---|
| Discovery | Complete | Market research report |
| Definition | Complete | MVP PRD, stack detection, context files |
| Design | Complete | 3 interactive component previews, 7-language microcopy pack |
| Planning | Complete | 68 frontend tasks, 28 backend tasks, dependency graphs |
| Quality & Analytics | Complete | 74 BDD test cases, 18 GTM tracking events |
| Wrap-up | In Progress | Scorecards, gap reports, learnings capture |
3 co-evolution cycles completed so far, tracking jaan-to from v3.15.2 through v3.19.0.
| Layer | Technology |
|---|---|
| Frontend | React 19, Next.js 15, TailwindCSS 4, Zustand 5 |
| Backend | Node.js 22, Fastify 5, REST + WebSocket (Socket.io) |
| Database | PostgreSQL 16 (Supabase), Redis 7 (Upstash), Typesense |
| AI | OpenAI SDK |
| Testing | Vitest 3, Playwright 1.50, Storybook 8 |
| Infrastructure | AWS, Docker, GitHub Actions |
jaanify/
├── vendor/jaan-to/ # Git submodule — the jaan-to plugin (read-only)
├── jaan-to/ # Plugin workspace — outputs, templates, context, docs
├── scorecards/ # Per-skill quality scorecards (updated each cycle)
├── gap-reports/ # Per-cycle scan reports and gap analyses
│ ├── 01-cycle/
│ ├── 02-cycle/
│ └── 03-cycle/
└── CLAUDE.md # Project rules (single source of truth)
# Clone with submodules
git clone --recurse-submodules https://github.com/parhumm/jaanify.git
cd jaanify
# Pull latest jaan-to plugin
git submodule update --remote --merge vendor/jaan-toTBD
TBD