Dypt

Dypt is an AI project builder tool. A task manager built around infinite task breakdown, where AI suggests subtasks, estimates time and turns notes or photos into structured to-dos. Deadlines, reminders, recurring tasks, dependencies, sharing and a CLI round it out, with generous hierarchy but tight free-plan limits.

Dypt is a task manager built around a single idea: work that feels too large to start becomes manageable once it is split into smaller pieces. Every task can be divided into subtasks, and those subtasks divided again, with no fixed limit on nesting depth. When the next step is unclear, the built-in AI proposes the missing subtasks, estimates how long work will take, or turns a photo of handwritten notes and whiteboards into structured tasks. Day to day, dypt behaves like a conventional to-do app with unusual depth. Tasks carry low, medium or high priorities, deadlines with push reminders, time estimates and actuals for tracking, markdown notes, reusable tags and note templates. Dependencies link tasks so blocked work is visible, and tasks can repeat on daily, weekly, monthly or yearly schedules. A calendar view and a list view cover planning, a command palette handles fast navigation, and weekly or daily email reports summarise progress. Sharing is a core feature rather than an afterthought. Tasks can be published publicly for anyone to view without an account, or shared privately with a set number of collaborators depending on plan, and the community library lets users copy public tasks into their own account and adapt them. A published command line interface extends the tool beyond the app, aimed at people and AI agents that work with structured task data and JSON output. Role-based starting points exist too, with an accountant workflow covering client work, month-end close, review notes and deadlines, and more roles listed as coming. Dypt sits between simple personal to-do apps and heavyweight project platforms. It is a reasonable fit for individuals and small teams who want hierarchy and light AI assistance, though anyone needing large free task volumes, per-seat pricing or a deep integration ecosystem will find the limits and modest collaboration caps restrictive.