EQueue AI

EQueue AI helps travelers find the best border queue time using natural language. AI explains recommendations while deterministic booking safely validates and reserves available slots.

Comment

EQueue AI

EQueue AI is an AI-assisted Telegram queue system designed to help travelers reserve a border crossing time using natural language.

Instead of manually comparing dozens of available time slots, users can simply describe their preference:

  • "Tomorrow after 10 AM"
  • "Earlier in the morning"
  • "Closer to noon"

The AI Planner interprets the request, while deterministic application logic validates availability, ranks eligible slots, and requires an explicit booking confirmation.

AI never allocates queue positions directly.

Inspiration

The original idea was created long before this GitHub repository existed.

The project initially lived only on my personal computer as a local prototype built to solve a real problem at the Narva–Ivangorod border crossing.

During OpenAI Build Week 2026 the project was significantly redesigned, documented, tested, and extended with an AI Planner powered by GPT-5.6 and Codex.

How it works

  1. User describes the desired time.
  2. GPT-5.6 converts the request into structured planning intent.
  3. The application validates business rules.
  4. PostgreSQL availability is checked.
  5. Eligible slots are ranked.
  6. AI explains the recommendation.
  7. The user explicitly selects a slot.
  8. The existing booking system safely creates the reservation.

Why AI

AI improves the user experience by understanding natural language and explaining recommendations.

Critical operations such as availability calculation, ranking validation, booking creation, and reservation safety remain deterministic inside the application.

Current status

Implemented:

  • Telegram booking flow
  • Canonical PostgreSQL availability
  • Atomic booking
  • AI Planner MVP
  • Secure booking handoff
  • Multi-language interface
  • Documentation and testing

Planned:

  • Live traffic information
  • Queue prediction
  • External data sources
  • AI Queue Companion
  • Web interface

Built With

Share this project:

Updates