Skip to content

vyom006/agentic_rag_travel_planner

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 

Repository files navigation

App

  • A full-stack AI-powered RAG-based app represented as a chat bot that lets the user plan their travel.
  • Users can get travel recommendations and check flight fares without leaving current UI
  • Unlike traditional apps where the user will have to click and fill-in multiple fields to provide their inputs, in this app, they can do all these things with 0 clicks and all by just chatting
  • This app is ai-powered in every front (from understanding user inputs to using search apis to using multiple ai agents)

Requirements:

  1. This is a near-real time* travel companion app
  2. Users will "chat" with the app as if they're talking to a person
  3. Users may to request for travel destination depending on their preference.
    • Users may have special occasions like summer vacation, wedding anniversary, birthday etc
    • The App will then consider user's preferences, special occasions, geo-political situations, weather situations
    • Finally, app will recommend travel destinations
  4. Next, users may ask "I want to know flights available between delhi and nyc on october 28th 2025"
  5. The App launches a web search using Perplexity's search api.
    • Get the top N results (say, N <= 8)
    • shows them in the results UI (say, in the right pane)
    • provides the summary of the result as audio also (openai's audio)
  • near-real time: as fresh as data available at Perplexity's search provider indexes

Tech:

  • Backend: Flask (Python)
  • Frontend: React (with vite and Typescript with bun)
  • LLM Models:
    • Reasoning (gpt-5-mini)
    • Intent + Summarization: gpt-4o-mini
    • Tools calling
  • AI Agents
  • RAG (Retrieval Augmented Generation)
  • Search API: Perplexity Search
  • Prompt Engineering
  • Vibe coding

Architecture:

Screenshot 2025-10-11 at 7 09 35 AM