About DeadlinePilot AI

An autonomous deadline monitoring agent that watches your tasks, reasons about urgency, and sends proactive reminders before work becomes panic.

What is DeadlinePilot AI?

DeadlinePilot AI is an autonomous deadline monitoring agent designed to reduce the cognitive load of keeping track of tasks. Instead of you checking dashboards repeatedly, the agent periodically scans your task list, evaluates urgency through AI reasoning, and proactively notifies you via Telegram and email before deadlines become emergencies.

How It Works

Client-Server Architecture

A Next.js frontend provides the user interface with App Router and React Server Components. A FastAPI backend handles all business logic, agent scheduling, and AI communication. The two communicate through a REST API.

REST API Communication

All data flows through a typed REST API: tasks, notifications, agent runs, and settings are exchanged as JSON between the frontend and backend. SSE streaming is used for real-time AI response visualization during plan generation.

AI-Powered Planning

The agent uses OpenAI-compatible models to generate structured task plans, evaluate urgency, and produce daily briefings. Responses are streamed via SSE so users see the AI thinking in real time.

Agentic AI Loop

The agent runs an autonomous Observe-Reason-Decide-Act-Memory cycle on a configurable schedule. Each run records its observations, decisions, and actions for full auditability.

O-R-D-A-M Pipeline

Step 1

Observe

The agent scans all active tasks, checks deadlines, progress, and priorities to build a current picture of your workload.

Step 2

Reason

Using AI, the agent evaluates which tasks are urgent, which are falling behind, and what the best next action is.

Step 3

Decide

Based on reasoning, the agent decides whether to send a notification, generate a briefing, or take no action.

Step 4

Act

The agent executes: sending Telegram messages, dispatching emails to Resend, or posting dashboard alerts.

Step 5

Memory

Every run is recorded. The agent remembers past states to avoid duplicate alerts and improve future decisions.

Tech Stack

Next.js

Frontend framework — React, App Router, SSR

FastAPI

Backend framework — Python, async, REST

SQLite

Embedded database — zero config, single file

OpenAI

AI reasoning — GPT / compatible models

Messaging — bot notifications

Resend

Transactional email delivery

APScheduler

Background job scheduling

Real-Time AI Response with SSE

When the AI generates a plan or a briefing, the frontend receives tokens as they are produced through Server-Sent Events (SSE). This streaming approach keeps the interface responsive, showing the AI thinking in real time rather than forcing users to wait for a complete response.