Inspiration

To help a travelling agencies to manage the multiple customer messages at same time. Manually it's very tough to provide relevant replies to customer and making them satisfied with the available packages.

What it does

Their is instagram page of that travelling agency called Anaghalaxmitours , you just have to send a message to that account and the bot will automatically activate and start chatting like a human.

How we built it

We built an automated Instagram messaging system where the Gemini API handles generating smart replies, while a Node.js backend manages the flow of messages. It connects to Instagram through the Instagram Graph API so it can receive DMs and respond instantly. The whole system runs on a cloud platform like AWS, and we used MongoDB to store chat data and keep track of conversations. With Express.js handling the API side, everything works together smoothly to create a reliable, real-time automation setup.

Challenges we ran into

One of the main challenges we ran into was working with the Instagram Graph API, since setting up permissions and webhooks wasn’t always straightforward and sometimes messages didn’t come through instantly. We also had to fine-tune how our Node.js backend handled multiple conversations at once so things wouldn’t slow down. Getting reliable and natural responses from the Gemini API took some trial and error, especially when dealing with different types of user inputs. On top of that, managing and storing chat data in MongoDB while keeping everything running smoothly on AWS required a bit of optimization. Overall, it was a mix of API limitations, performance tuning, and improving response quality that made the process challenging.

Accomplishments that we're proud of

We are proud to have built an automated Instagram DM system that helps a travel agency handle multiple customer conversations at the same time without needing constant manual replies. By integrating the Instagram Graph API with a Node.js backend, we enabled real-time message handling and instant responses, which greatly improved customer communication speed. The Gemini API helps generate natural, human-like replies so customers feel like they are talking to a real person. We also used AWS and MongoDB to make the system scalable and reliable as the number of conversations grows. Even though we faced challenges with API setup, webhooks, and managing multiple chats smoothly, we were able to build a stable system that reduces a lot of manual work for travel agents and lets them focus more on helping customers plan and book their trips.

What we learned

Here’s a natural, human-written “What we learned” paragraph you can use:


Through this project, we learned how to build and manage a real-time messaging automation system and gained hands-on experience with integrating multiple technologies together. Working with the Instagram Graph API taught us how important proper setup, permissions, and webhook handling are for reliable communication. We also improved our backend development skills using Node.js and Express, especially in handling multiple user conversations at the same time without losing performance. Using the Gemini API helped us understand how to fine-tune prompts to generate more natural and context-aware responses. On the infrastructure side, deploying on AWS and managing data with MongoDB gave us practical exposure to building scalable systems. Overall, this project helped us understand how different services work together in a production-like environment and how small technical issues can impact the overall user experience.

What's next for Travelling agent consulatant (chat bot)

Here’s a natural, human-written “What’s next” section you can use:


Next, we plan to make the travel chatbot more intelligent and useful for both customers and travel agents. One major improvement is adding personalized recommendations, where the bot suggests travel packages based on user budget, preferences, and past conversations. We also aim to integrate payment links and booking confirmation directly inside the chat so users can complete their bookings without leaving Instagram. Another focus is improving multilingual support so it can comfortably interact with customers in different languages. On the technical side, we want to add better analytics for travel agents, showing insights like most asked destinations, conversion rates, and peak inquiry times. We also plan to improve the AI responses further so they feel even more natural and context-aware, especially for complex travel planning questions. Overall, the goal is to turn it into a complete AI travel assistant that not only answers queries but also helps close bookings efficiently.

Share this project:

Updates