Boscof AI | Chatbot that amplifies a creator's voice
All of Álvaro Boscof's knowledge poured into an AI. A tool for his community to access his philosophy anytime, anywhere.
Context
Álvaro Boscof is a personal development and mindset creator with over 1 million followers. The project started as a personal idea and ended up becoming a direct collaboration with him.
Goal
Build a chatbot able to answer questions in Boscof's own style, tone and philosophy, so his community could access his knowledge without him needing to be present in every interaction.
Todo empezó con un script de Python
Llevaba tiempo siguiendo a Álvaro Boscof: uno de los referentes más conocidos del desarrollo personal y la mentalidad en español, con más de 1 millón de seguidores. Un día me acordé de que tenía un script en Python que transcribía vídeos de YouTube y audios. Lo puse a funcionar, transcribí todos sus vídeos públicos y con ese material creé un GPT personalizado entrenado con su contenido.
La idea era solo para mí. Quería poder preguntarle cosas a Boscof sin tener que buscar en qué vídeo lo había dicho.
Lo que pasó después
Le envié el GPT. Al rato me respondió: llevaba tiempo pensando en crear un chatbot que hablara como él, y quería saber si yo podía hacerlo.
Le dije que sí. Siendo completamente honesto, en ese momento no sabía exactamente por dónde empezar.
Un mes de investigación y resolución de problemas
Lo que vino después fue un mes de trabajo real: investigar, probar, romper cosas, volver a empezar. Cada problema que aparecía lo iba resolviendo sobre la marcha. Sin hoja de ruta clara, sin garantías de que fuera a salir.
El resultado fue un chatbot capaz de responder preguntas con el tono, el estilo y la filosofía de Boscof, entrenado con todo su contenido público. Una herramienta que no solo automatiza respuestas, sino que mantiene su forma de pensar y de comunicar. Su conocimiento disponible 24/7, sin que él tenga que estar presente en cada conversación.
Lo que me llevo
Este proyecto me enseñó que los mejores proyectos a veces no se planifican: ocurren. Y que decir que sí cuando no sabes del todo cómo hacerlo, si tienes curiosidad y constancia, casi siempre sale adelante.
It all started with a Python script
I’d been following Álvaro Boscof for a while: one of the most well-known personal development and mindset creators in Spanish, with over 1 million followers. One day I remembered I had a Python script that transcribed YouTube videos and audio files. I ran it, transcribed all his public videos, and used that material to create a custom GPT trained on his content.
The idea was just for me. I wanted to be able to ask Boscof things without having to dig through dozens of videos to find where he’d said it.
What happened next
I sent him the GPT. He replied shortly after: he’d been thinking for a while about creating a chatbot that talked like him, and he wanted to know if I could build it.
I said yes. Being completely honest, at that point I didn’t know exactly where to start.
A month of research and problem-solving
What followed was a month of real work: researching, testing, breaking things, starting over. Every problem that came up, I worked through it as it appeared. No clear roadmap, no guarantee it would work out.
The result was a chatbot capable of answering questions in Boscof’s tone, style and philosophy, trained on all his public content. A tool that doesn’t just automate responses — it maintains his way of thinking and communicating. His knowledge available 24/7, without him needing to be present in every conversation.
What I take away
This project taught me that the best projects sometimes aren’t planned: they just happen. And that saying yes when you don’t fully know how to do something — if you have curiosity and persistence — almost always works out.
Lessons learned
A personal project with no intention of showing it can open doors you didn't expect. Sometimes the most valuable thing is what you build for yourself.
Not knowing how to start something isn't a reason not to try. Research and persistence end up solving almost any technical problem.
AI doesn't replace a creator's essence, it amplifies it. That was the hypothesis and the result confirmed it.