As messengers have become one of the most used apps both on smartphones and desktop devices, chatbots have also gained popularity, as they allow people to do various different things without leaving their messenger app. Chatbots can provide many services, such as customer support, online shopping and marketing. However, creating a chatbot requires coding experience or hiring a specialist. Chatforma solves that issue, giving people an easy and functional chatbot constructor, that supports many platforms such as Telegram, Facebook Messenger, Viber and VK.

The goal

Initially the client already had a working version of the platform with real users. My goal was to develop new features and fix some of the bugs the current platform had. However, due to poor architecture creating new features was financially unprofitable because of the issues the current codebase had. I estimated that in about 2-3 months the project could be rewritten from scratch with all the old and new features, but with the right architecture the further development and maintanance would become much faster and easier. The client agreed to our proposal and my team started working on the new version of the platform.

My role

At the start of the project I was invited as a React.js and Node.js developer. My responsibility was to audit the current codebase and offer improvements, developing new features and fixing bugs alongside. After the client agreed to rewrite the project from scratch, I took the role of the team lead and led the development of Chatforma v2. I was managing the design of the platform’s UI, creating prototypes and working on the UX. After the design of the UI was made, I’ve designed the architecture of the backend and frontend of the project as a fullstack javascript developer and also managed 2 other developers and thier work.

Message sending

Technologies

Originally the project was created with a cloud-backend solution based on Node.js, that was later deprecated and taken down, which forced the developers to run the server on their own. MongoDB was used as a database and React.js SPA for the frontend.

Сloud backend was replaced with a Node.js API using Adonis.js framework, which provides all of the features needed to create a powerful Restful API, has excellent documentation and is easily extendable. The MongoDB was also replaced with PostgreSQL as it was more important to have a stable and reliable schema.

The fronted was written with a battle-tested React.js stack and consisted mainly of complex forms and fields.

The final stack for the project was:

• React.js + Redux + styled-components for frontend
• Node.js with Adonis.js framework for both dashboard API and chatbot’s backend with unit tests
• Redis for message distribution queue
• PostgreSQL as a database

Form data collection

Result

The project was successfully rewritten, introducing new features and fixing countless amount of bugs. The UX was drastically improved alongside the performance of the connected chatbots. All the chatbots linked to the first version of the platform were fully migrated, so that users won’t loose anything during the transition to the new version. And most importanly, new architecture allowed to release new features multiple times faster, making clients happy and allowing the business to grow.