Zero-Cost, AI-Native Tech Blog powered by Local LLM Pipeline
Live Site: https://hun-bot.dev
Hybrid Architecture: Serverless Frontend (Vercel) + Local AI Backend (Mac M4 Pro).
| Component | Tech Stack | Description |
|---|---|---|
| Orchestrator | Python 3.11 | Automated batch translation pipeline using tqdm & psutil. |
| Inference | Ollama | Zero-Cost local serving (En: gemma2:9b / Jp: qwen2.5:7b). |
| Observability | Grafana + PG | Real-time monitoring of TPS (Tokens/sec) & GFLOPS. |
| Consistency | Regex/YAML | Strict metadata parsing to prevent structure collapse. |