Home Creators Posts Import Register Favorites Logout

Content

If you’ve ever wanted to build your own AI anime companion but don’t know where to begin, this roadmap breaks the process into clear, simple steps.

Here’s what we’ll cover in the video:

Foundations: Python basics, version control (Git + VS Code), and getting comfortable with the command line.

APIs and LLMs: How to set up an API key, send your first request to an LLM, and handle responses.

Speech Tools: Starting with text-to-speech (TTS) and speech-to-text (STT).

Local vs Cloud: When to use paid APIs and when it makes sense to run models locally.

Going Further: An introduction to fine-tuning local models, and the basics of connecting front-end, back-end, and client-side systems together.

I’ve attached the roadmap as a PNG you can follow along with during the video. This one is completely free, so feel free to use the roadmap at your own pace or alongside the walkthrough.

Files

Comments

Sanryusan

Thanks for the video 😄

GLuz

My goal is quite ambitious: I want to create a waifu that help me unfuck my life, organize my medicine, calm me down, create reminders and operate my homeassistant and my desktop. That's why I'm building ~hanauta~for that matter so that I'll integrate i3 or bspwm desktop with productivity tools and an AI avatar

john

Im following this in hopes to finish it and make one myself as part of my computer science class. I have to present a project and i think this could be a awsome way to do it

Nehal

how much did you improved so far ?

Gowtham Kodali

I am going to be honest he did a really good job getting a person into AI and ML without advanced topics like how a AI works and how to train your own. I would say that most of the time you will use claude or any other AI to help you code cause it makes it faster but you will need to understand EVERYTHING that he talks about to actually test or even understand basic terms. It took me a good half a year to learn these terms by myself and test by myself, I really appreciate Rayen's work ;)

Fewzan Mohmud

What if I go in my own pace like using avatars early or use local llms and form a workflow early as well because in my recent project Iam trying to have it be able to be interact with voice and also integrate ai workflow