Skip to content
View Shagwu's full-sized avatar

Block or report Shagwu

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Shagwu/README.md

I build AI systems and document the whole messy process.

Indie AI builder. Content creator. Practitioner first, teacher second.

I'm building Road4AI — an open-source, zero-cost blueprint for operating as a solo AI builder using a real multi-agent stack. No paid APIs required. No bloated frameworks. Just a working system you can actually clone, run, and adapt.

Everything I build gets documented in public — the wins, the broken builds, the decisions I had to make twice.


What I'm building

Road4AI

A layered multi-agent system for solo operators. CLI-native, local-first, built on tools that don't cost money to run.

The stack: Gemini CLI for ideation and research · Codex for structured builds · Claude for reasoning and strategy · Hermes v2.0 as the distributed memory layer that ties it all together.

The idea: one person, a set of agents that actually talk to each other, and a system that doesn't require a startup budget to run.

Currently hardening Hermes v2.0 — the memory backbone that lets any agent in the system pick up where the last one left off, across sessions, across tools, across context resets.

→ See the repo · → See it live


How I work

Every agent in Road4AI operates under a shared contract — what to build, how to behave, when to stop and ask. The system is designed to be lean at every stage. Habits before infrastructure. One validated thing at a time.

I document the build in real time across five platforms. Not polished tutorials. Actual practitioner updates — what I tried, what broke, what I'd do differently.


The stack

Layer Tool What it does
Ideation + research Gemini CLI First-pass thinking, fast iteration
Structured builds Codex Code, refactoring, precise edits
Reasoning + strategy Claude Long-context thinking, content strategy
Memory Hermes v2.0 Distributed context across all agents
Distribution Blotato Scheduled social output
Git workflow GitNexus Version control tooling

Follow the build

I post the real-time journey across five platforms — pick whichever fits how you read.

  • LinkedIn — practitioner updates, what I'm learning
  • Instagram@road4ai — the build in short-form video
  • Threads — punchy takes, quick decisions
  • TikTok — same as Instagram, for the algorithm
  • X — direct, no filler

Part of something bigger

Road4AI started as my own operating system. It's becoming a reference others can use.

If you're a woman building with AI — or just a solo builder trying to figure out how multi-agent systems actually work in practice — this is built for you. I'm part of a Skool community of women building with AI, and everything I validate here gets shared there first.


May 2026 · Hermes v2.0 reveal incoming.

Popular repositories Loading

  1. Shagwu Shagwu Public

    Config files for my GitHub profile.

  2. exclusive exclusive Public

    My clone repository

  3. mighty2 mighty2 Public

    Fantastic-adventure

    Python

  4. deepseek-api-shagwu deepseek-api-shagwu Public

  5. agents-course-summary agents-course-summary Public

    MDX

  6. agents-course-summary1 agents-course-summary1 Public

    Hugging Face Course Summary