About the Project

agisoul is an independent AGI lab: a personal project about experiments with LLM agents and the philosophy of artificial intelligence. The project is run by a single author; it is not a scientific organization or a company.

Why you can trust

The project rule: minimal live example first, then text. Every piece of code in the courses and blog is run and verified, not rewritten from memory. Mistakes still happen — I show them along with solutions: the cases of the services subproject are structured this way.

For whom

Just to use — analyses without code: how models are arranged and what to adopt in everyday life: Разборы
Beginner programmers — course “Agent from Scratch”: from the agent formula to your own CLI agent
Small and medium-sized businesses — automation cases in the services subproject (as they accumulate) and analyses of “which neural network to choose for the task”

Questions — in Telegram (✈ icon next to sections in the sidebar), I answer myself.

What’s inside

Wiki — 130+ atomic notes on Li Bozhe’s book “Deep Understanding of AI Agent”: one note = one idea, all linked into a single network
Learning — mini-course “Agent from Scratch”: from the agent formula to your own CLI agent
Blog — laboratory journal: analyses and digests; applied cases — in the services subproject
Telegram — short signals and updates (✈ icon next to sections in the sidebar)

What’s not here

Statements about creating AGI and guarantees of implementation results — the results of experiments depend on the models, data, and usage conditions.

Materials are prepared with the help of LLM and verified by the author: errors are possible, important things should be checked against the original sources. Publications are for informational purposes only and do not constitute professional advice.