Skip to content
View avenoxai's full-sized avatar
🥰
waiting4pdoom
🥰
waiting4pdoom

Block or report avenoxai

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.

Content in all repositories owned by your account will be closed.
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
avenoxai/README.md

Avenox. AI systems, built in the open. The Limon mascot at the drafting table.

Website   /   YouTube   /   X   /   Get in touch


I’m Taha, a.k.a. Avenox. I build agent systems, test models on real tasks, and share the tools behind my work.

On YouTube, I walk through the process in Turkish. Here, you can explore the code, reuse the skills, and inspect the experiments for yourself.

Start here

01 / MEMORY

An open-source second brain built around Obsidian and Claude Code. Keep useful context across sessions.

Build your second brain →

02 / WORKFLOWS

Reusable agent skills from my day-to-day work: coding fleets, reviews, research, and video production.

Explore the skills →

03 / EXPERIMENTS

Raw frontier-model experiments from the videos. Original prompts and outputs, ready to inspect.

See what the models made →

04 / WATCH THE PROCESS

Hands-on builds, model comparisons, and the practical work of turning AI tools into useful systems.

Watch the builds →

On the workbench: Serai

A shared control plane for AI agents.

I’m building Serai to connect memory, context, tools, approvals, and execution records across agent workflows. The aim is continuity: useful work should carry across sessions, tools, and teams.

In active development. The codebase is currently private.

Explore Serai · Open the hub

A few experiments to open

Experiment What to look for
AI Ne Kadar İlerledi? Models from different generations tackling the same visualization tasks.
Sonnet 5 vs Opus 4.8 Identical prompts, with both models’ original outputs.
GPT-5.6 SOL Benchmark Single-file demos and the prompts that produced them.
Galaxy simulation A browser-based Milky Way experiment.


Build. Test. Inspect. Share.
For collaborations and sponsorships: taha@avenox.lol

Pinned Loading

  1. sifirdan sifirdan Public

    Yapay zekaya sıfırdan başlayanlar için haftalık seri — notlar, sözlük ve bölüm siteleri

    CSS 134 8

  2. avenoxbeyin avenoxbeyin Public

    Your AI second brain in one command — open-source Obsidian + Claude Code system with persistent cross-session memory. avenox.lol/beyin.md

    Python 183 38

  3. avenoxskills avenoxskills Public

    Production agent skills for Claude Code, Cursor, and any SKILL.md harness — Codex fleets, video pipeline, monorepo review bundles, multi-chain explorer.

    Python 73 6

  4. avenox-hermes-notlari avenox-hermes-notlari Public

    Kendi sunucunda 7/24 çalışan bir AI operatörü kurmanın desenleri — Hermes Agent videosunun notları

    25 2

  5. avenoxstatusline avenoxstatusline Public

    Two-line Claude Code status line with a state-driven ASCII pet. 132 lines of bash, no npm.

    Shell 16 5

  6. higgsfield-studio higgsfield-studio Public

    Claude Code + Higgsfield ile görsel/video üretim skill'leri — GPT Image 2, Seedance loop'ları, sunum sitesi

    9