$ fastfetch --profile samir
samir@github
──────────────────────────────────────────────────────
Name Samir Hanna Verza
Host Brazil
Companies Blue3 · B3dev
Role full-stack builder · AI/LLM engineering
Kernel Rust · Ruby · PHP · Java · TypeScript · Python
Shell bash · anna (agentic CLI I built)
DE GNOME (with my own extensions)
Surfaces web · desktop · mobile · terminal
Uptime always building
graph LR
ME(("@samirhvbr")) --> B["🏢 Blue3 · B3dev"]
ME --> T["🖥️ SShvTerm<br/>SSH terminal with AI"]
ME --> OSS["🌱 open source"]
B --> S["🧠 SHVIA<br/>AI platform"]
B --> E["📊 EOP<br/>platform for ISPs"]
S --> W["🌐 web"]
S --> D["🖥️ desktop · Tauri"]
S --> M["📱 mobile · iOS/Android"]
S --> A["⌨️ anna · agentic CLI in Rust"]
OSS --> U["ai-usagebar<br/>GNOME · macOS · Windows"]
OSS --> LEB["AI-BENCHMARK<br/>LLMs on legacy code"]
| project | what it is |
|---|---|
| SHVIA | Multi-surface AI platform — web, desktop (Tauri), mobile iOS/Android and CLI. Multi-provider BYOK gateway, voice (STT/TTS) and code mode. Born at Blue3 / B3dev. |
| anna | Agentic CLI in Rust — tool loop with safety gates, embeddable in other apps. It is SHVIA's code engine. |
| SShvTerm | SSH terminal (Tauri 2 desktop + mobile) with an embedded AI agent — the terminal that talks back. · sshvterm.com |
| EOP | Enterprise Operating Platform — ledger-first management for internet service providers, at Blue3. |
- AI-BENCHMARK — RFC + harness for the LLM Engineering Benchmark: measures LLMs fixing legacy code without breaking compatibility — the opposite of greenfield
- ai-usagebar — desktop integrations (GNOME · macOS · Windows) contributed upstream
- GITHUB-DESKTOP — GitHub Desktop fork with a multi-repo dashboard
- hermes-agent — fork of Nous Research's open agent
- Vitals · claude-desktop-debian · Matomo · CSL-Redes — day-to-day forks and contributions
[languages]
daily = ["Rust", "Ruby", "PHP", "TypeScript", "Java", "Python", "Swift"]
[frameworks]
apps = ["Tauri 2", "Rails", "Laravel", "Spring Boot", "React + Vite"]
[data]
databases = ["PostgreSQL", "PostGIS", "pgvector", "MariaDB"]
[ai]
tracks = ["inference gateways", "agents", "RAG", "STT/TTS", "LLM benchmarking"]
[surfaces]
targets = ["web", "desktop Linux/macOS/Windows", "mobile iOS/Android", "CLI", "GNOME Shell"] 998 ship the SHVIA mobile client to the app stores
999 contribute the ai-usagebar desktop integrations upstream
1000 spec out the LLM Engineering Benchmark — AI on legacy code
- 🌐 samirhv.com.br — personal page & download hub
- ✍️ AREA81 — the blog
- 🖥️ sshvterm.com — SShvTerm
- 🏢 blue3.com.br · b3dev.com.br
- ✉️ samirhv@proton.me
$ exit — the prompt stays open · from backend to binary 🇧🇷