MindFlow is a full-stack web platform built with Next.js, NestJS, and PostgreSQL. It uses an EMA (Ecological Momentary Assessment) chatbot to measure a student’s mental fatigue on a 1–5 scale, then automatically breaks down difficult tasks into micro-objectives of 25 minutes or less — reducing cognitive friction and making it easy to take the next step.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/J0S3-LEON/pf_pruebasAseguramientoCalid_SDD/llms.txt
Use this file to discover all available pages before exploring further.
Introduction
Understand the problem MindFlow solves and how the system fits together.
Quickstart
Clone the repo, set environment variables, and run MindFlow locally in minutes.
Architecture
Explore the three-tier monorepo: Next.js frontend, NestJS API, and PostgreSQL.
API Reference
Full reference for every REST endpoint under
/api/v1/.How MindFlow Works
Student logs in
Register with an email and password. MindFlow issues a signed JWT that grants access to all protected endpoints for 24 hours.
Start an EMA session
The EMA chatbot asks for a fatigue score from 1 to 5. This score is persisted with a UTC timestamp and linked to the active session.
Tasks get intelligently adapted
If fatigue ≥ 4, the Task Decomposer calls an external LLM to split every active task into 2–7 micro-objectives, each capped at 25 minutes. If fatigue ≤ 3, tasks are shown in their original form.
Key Features
EMA Chatbot
Real-time fatigue assessment via a conversational interface. Scores outside [1, 5] are rejected without being recorded.
Task Decomposer
AI-powered breakdown of complex tasks into actionable micro-objectives when fatigue is high.
Student Dashboard
Consolidated view of tasks, micro-objectives, and a 30-session fatigue history chart.
Notifications
Contextual reminders with frequency limits (≤3/day) and suppression during active EMA sessions.
Authentication
JWT-based auth with bcrypt hashing, 24-hour token expiry, and global guard with public route exceptions.
Property-Based Testing
27 correctness properties verified with fast-check across hundreds of random inputs per run.