VELOX

A easy way to find street racing events happening near you.

My Journey and the Birth of Velox

I’ve always enjoyed running and taking part in local street races. Even though I’m not the fastest, I’ve always loved being part of it.
Last year, I decided to challenge myself — not only by running my first half marathon but also by creating a running group with my friends: the Dash Running Club.

During that time, I noticed a problem. There were many companies organizing races, but the information was scattered across platforms, websites, and Instagram pages.
We had to manually share the main races, dates, locations, and registration links in our WhatsApp group.
That’s when the idea of automating this process came up.

Velox was born from the need to centralize and organize all running-related information in one place.
We don’t create or sponsor events (yet). Our goal is to make runners’ lives easier by collecting and organizing race data across São Paulo.

Today, we offer our online platform, our race calendar, and our Telegram bot.

Features

Online Calendar

We list races across São Paulo and the surrounding region in a single, easy-to-use calendar. You can find all the details about your next race without searching through multiple websites or organizers.

Telegram Bot

Our bot is available 24/7 to assist runners.
It can find races, list event details, share Google Maps locations, race times, and even the weather forecast.
It also works inside private Telegram groups without extra setup — just add it, and it runs smoothly.

AI Chat

After years of working in Customer Experience (CX), I built a custom chat system for Velox.
It can handle complex questions about the platform, events, and public information, helping users find what they need effortlessly.

Technical Overview

Velox is built on three main services that communicate through well-defined contracts: the Back-end, the Front-end, and the Chatbot.
Each service is deployed independently but follows shared standards for type safety, automation, and consistency.
The architecture emphasizes modularity, domain isolation, and observability — key principles that support scalability and reliability.

Architecture explanations

1 - Why 2 Front-ends?

For me, it doesn’t make sense to have both the user interface and the admin dashboard in the same front-end project. If I did that, I’d need a complex and highly reliable authentication system, which would add unnecessary complexity to a simple user-facing application.

Of course, my management tool includes login and protection layers, but since it’s an internal tool used only by me, keeping it separate was a deliberate design choice for the initial scope of the project.

2 - Using OpenAI directly on the Front-end/Bot

I use direct integration with OpenAI on both the front-end and the bot. This allows the system to understand user messages naturally — in both the bot interactions and the embedded web chat.

This approach reduces response time by avoiding unnecessary calls to the back-end. When dealing with a chat-based tool, response speed is crucial for a good experience.

To prevent abuse or security issues, both projects include protection layers — such as input sanitization, structured prompts, and strict rate limiting.

Back-end

The back-end manages all business logic and data persistence.
It follows a Clean Architecture approach with well-structured layers and explicit interfaces.
Each endpoint handles authorization, validation, and response formatting in a consistent pattern to ensure security and predictability.
Data validation and monitoring tools ensure input safety, while observability is built in to measure performance, errors, and user activity.

Front-end

The front-end transforms raw data into a clear, engaging experience for both runners and administrators.
It uses a modern React and Next.js stack with hybrid rendering (SSR + SSG) and reusable components across public and internal views.
The design system ensures accessibility, visual consistency, and smooth animations, while telemetry tools capture engagement and user behavior metrics.
The platform is designed for omnichannel use — from public pages and event listings to dashboards that track participation and insights.

Chatbot

The chatbot follows a modular, Hexagonal Architecture pattern to keep core logic independent of the delivery channel.
It supports multiple commands such as /start, /help, and /races, and integrates with AI to understand natural language inputs that don’t match predefined commands.
When users send unrecognized messages, the bot interprets intent and responds intelligently — either suggesting the right command or delivering a helpful answer.
This architecture also enables expansion to future platforms like WhatsApp, Discord, or Slack without major code changes.

More than i ideia

Velox started as a small side project meant to make my life easier.
Over time, I realized it was much bigger than I first imagined — something that could actually help many more people than I expected.
It’s not the first project I’ve brought to life, but it’s the one that truly has the potential to grow.

Business Model

Currently, there’s no platform in Brazil — or elsewhere — that centralizes running events the way Velox does.
The project started as a personal initiative and is now sustained through community support and early-stage partnerships with local brands and event organizers.

As our user base grows, we plan to introduce targeted marketing opportunities for sponsors in the running and sports ecosystem, ensuring the project remains sustainable and community-driven.

Thanks to free and open-source resources, our operating costs remain minimal — roughly $15 per month is enough to keep everything running smoothly.