# BASHCAT.NET

> BASHCAT.NET 是台灣的 AI 與嵌入式系統開發團隊（成立於 2020），專精單晶片（8051 / STM32 / ESP32 / Arduino / Nordic nRF）韌體開發、PCB 設計、AI 與機器學習應用（RAG、Computer Vision、Edge AI）、Web 與雲端系統整合。完成 50+ 商業專案，平均縮短 40% 開發週期。提供全台灣現場與全球遠端服務。

This file follows the [llms.txt](https://llmstxt.org) standard so generative engines (ChatGPT, Claude, Perplexity, Gemini, Apple Intelligence) can find a canonical map of BASHCAT.NET content.

For full machine-readable content, see [llms-full.txt](https://bashcat.net/llms-full.txt).

## Core pages

- [Home / 首頁](https://bashcat.net): What BASHCAT.NET does, hero stats, marquee tech stack, services preview, products preview.
- [About / 關於](https://bashcat.net/about): Team background, founding (2020), specialty areas, contact.
- [Services / 服務](https://bashcat.net/services): AI/ML, Embedded Systems, Web/Cloud — what we deliver and how.
- [Portfolio / 開發案例](https://bashcat.net/portfolio): 12 productized case studies across AI, IoT, Embedded, Web.
- [Blog / 技術專欄](https://bashcat.net/blog): Long-form technical writing on STM32, ESP32, BLE, RAG, Edge AI, low-power design.
- [Contact / 聯絡](https://bashcat.net/contact): Email bashcat0804@gmail.com, free initial consultation.

## Products (own SaaS / native apps)

- [Saybit](https://bashcat.net/saybit): macOS menu-bar AI voice input. Semantic understanding, context awareness, developer mode, BYO API key. Free, v1.4.2.
- [LM 備忘錄 (LM Memo)](https://bashcat.net/lm-memo): iOS/iPadOS/watchOS/macOS AI voice memo with on-device WhisperKit + AI summarisation + auto-todos. NT$790, on the App Store.
- [LLM Cluster](https://bashcat.net/llm_cluster): Self-hosted distributed GPU sharing platform, OpenAI v1 compatible API, supports LM Studio + Ollama backends, dual load balancing. Beta, v0.8.3.
- [Muse Mail](https://bashcat.net/muse-mail): iOS AI mail assistant with summarisation, smart compose, tracking-pixel blocking, BYO LLM provider. Pre-order.

## Featured portfolio case studies

- [AI 智慧寵物互動攝影機](https://bashcat.net/portfolio/ai-pet-camera): Edge inference YOLOv8 + ESP32-S3 + WebRTC pet tracker.
- [Nordic nRF54L15 藍芽 6.0 開發實戰](https://bashcat.net/blog/nordic-nrf54l15-bluetooth-development): Bluetooth 6.0 + Channel Sounding 室內定位.
- [ECG 心率藍牙監測裝置](https://bashcat.net/portfolio/ecg-heart-rate-bluetooth-device): STM32 + BLE 5.0 medical-grade ECG.
- [LoRaWAN 農業環境監測系統](https://bashcat.net/portfolio/lorawan-agricultural-sensor): Solar-powered STM32L4 + TTN + Grafana.
- [企業 RAG 知識庫系統](https://bashcat.net/portfolio/rag-knowledge-base): On-prem LangChain + pgvector + GPT-4 + Next.js.
- [企業級數據分析平台](https://bashcat.net/portfolio/enterprise-data-analytics): Next.js + ClickHouse + Kafka + D3.js, millions of events per second.
- [預測性維護系統](https://bashcat.net/portfolio/predictive-maintenance-system): TinyML 工業 4.0 振動監測.
- [智慧家居自動化](https://bashcat.net/portfolio/smart-home-automation-system): Matter / Thread / Zigbee 整合.
- [藍牙室內定位](https://bashcat.net/portfolio/bluetooth-indoor-positioning): BLE 6.0 Channel Sounding 公分級定位.

## Technical blog posts — AI / LLM / Claude Code (2025-2026 熱門)

- [發現兩個 Claude Code 監控神器](https://bashcat.net/blog/claude-code-monitor-tools): ccusage、vibe-log token 用量分析（716k+ views）.
- [Claude Desktop 與 MCP 完整使用指南](https://bashcat.net/blog/claude-desktop-mcp-guide): Model Context Protocol 從零到實戰（614k+ views）.
- [Claude Code Sub Agents 完全指南](https://bashcat.net/blog/claude-code-sub-agents-guide): 從單兵作戰到 AI 軍團指揮官（611k+ views）.
- [Harness Engineering 完全解析](https://bashcat.net/blog/harness-engineering-ai-agents): AI Agent 護城河不是模型而是環境（513k+ views）.
- [Whisper.cpp macOS 完整安裝指南](https://bashcat.net/blog/whisper-cpp-macos-install): C++ 版 Whisper，Metal 加速本地轉錄.
- [Gemma 4 使用場景完全解析](https://bashcat.net/blog/gemma-4-use-cases-guide): 六大最強場景與避坑指南.
- [Mac 跑 AI Coding Agent — oMLX](https://bashcat.net/blog/omlx-mac-ai-coding-agent): MLX + SSD swap 突破 VRAM 限制.
- [Whisper 語音識別優化之路](https://bashcat.net/blog/whisper-asr-optimization): 中文轉錄品質從亂碼到完美的完整心法.
- [AI Agent 三國大亂鬥](https://bashcat.net/blog/ai-agent-comparison-2025): Google Jules vs OpenAI Codex vs Claude Code 評比.
- [AI 浪潮中活下來的繪師做對了什麼](https://bashcat.net/blog/artists-surviving-ai-era): AI 時代創作者的策略分析.

## Technical blog posts — 嵌入式 / 硬體 / IoT

- [Nordic nRF54L15 藍芽 6.0 開發實戰](https://bashcat.net/blog/nordic-nrf54l15-bluetooth-development): Channel Sounding / 室內定位 / Matter.
- [BLE 天線 Pi 型匹配 + NanoVNA 調試](https://bashcat.net/blog/ble-antenna-matching-nanovna): Nordic nRF52 RF 信號路徑與實戰匹配.
- [Nordic nRF Connect SDK Bare Metal 深度解析](https://bashcat.net/blog/nordic-nrf-baremetal-analysis): 不綁 Zephyr 的官方方案.
- [小訊號 PCB 設計指南](https://bashcat.net/blog/small-signal-pcb-design-guide): trace 寬度、屏蔽、地平面分割、差動對佈線.
- [KiCad Python API (kipy) 快速入門](https://bashcat.net/blog/kicad-python-api-kipy-intro): KiCad 9 原生 Python 自動化 PCB 工作流.
- [Raspberry Pico + Zephyr 開發](https://bashcat.net/blog/raspberry-pico-zephyr-dev): RP2040 跳脫 Pico SDK 的 RTOS 選項.
- [Meshtastic 使用場景全解析](https://bashcat.net/blog/meshtastic-use-cases): 登山 / 災害通訊 / 社區韌性 LoRa 網狀網路.
- [ESP32 IoT 實戰指南](https://bashcat.net/blog/esp32-iot-practical-guide): 從韌體到雲端 MQTT 全流程.
- [低功耗嵌入式設計](https://bashcat.net/blog/low-power-embedded-design): Sleep mode、能耗分析、電池壽命延長.
- [MQTT vs HTTP for IoT](https://bashcat.net/blog/mqtt-vs-http-iot): 何時該選哪個協定的決策樹.

## Technical blog posts — AI / Web / 跨領域

- [TinyML 邊緣 AI 指南](https://bashcat.net/blog/tinyml-edge-ai-guide): MCU 上跑模型、量化、TFLite Micro.
- [RAG 系統實作指南](https://bashcat.net/blog/rag-system-implementation-guide): 向量資料庫 + chunking 策略 + 引用回溯.
- [AI 整合最佳實踐](https://bashcat.net/blog/ai-integration-best-practices): 把 LLM 變成產品功能.
- [Flutter vs React Native 2024](https://bashcat.net/blog/flutter-vs-react-native-2024): 跨平台 App 選型決策.
- [學生 IoT 專題指南](https://bashcat.net/blog/student-iot-project-guide): 畢專/競賽從零到 demo 的完整路徑.

## Optional

- [English version](https://bashcat.net/en): All content available in English.
- [Sitemap XML](https://bashcat.net/sitemap.xml)
- [Contact email](mailto:bashcat0804@gmail.com)
- [GitHub: Oliver0804](https://github.com/Oliver0804)
- [YouTube: @bashcat987](https://www.youtube.com/@bashcat987)