Blog AI & HCI
ReAct Research Agent with n8n
주제 하나만 던지면, AI가 스스로 검색하고 판단하고 — 부족하면 다시 검색한다.
Built a ReAct (Reasoning + Acting) agent that autonomously repeats search and judgment until it has enough information to answer a research topic. Unlike a single-pass pipeline, this agent decides for itself whether to search again or stop, using Gemini for judgment and Tavily for …
AIRead more →
Claude Code Practical Guide
챗봇이랑 뭐가 달라요? 물어보는 사람이 많아서 정리했다.
A practical reference guide for Claude Code — the terminal-based AI coding agent by Anthropic. Covers the key difference from the chat interface, essential commands, and common gotchas that trip up first-time users.
Core concept: Claude Code runs in your terminal and acts directly on …
AIRead more →
OpenClaw: Open Source AI
오래된 맥북에 설치하고 텔레그램으로 말 걸었더니, 자비스가 됐다.
Tested OpenClaw — a free, open-source personal AI agent — by installing it on an old MacBook (Retina 2017) as a dedicated agent machine, instead of running it on my main laptop.
OpenClaw runs locally and connects to messaging apps like Telegram, WhatsApp, and Discord. …
AgentRead more →
Claude Cowork: Daily Brief
매일 아침 6시, AI가 영어 단어·뉴스·일정을 정리해서 메일로 보낸다 — 사람 손 없이.
Built a personal AI Agent that runs every morning at 6:00 AM and delivers a fully assembled daily brief to my inbox — without any manual trigger.
Architecture: Cowork acts as the local automation runtime. At 6:00 AM, a scheduled trigger fires Claude Sonnet with …
AgentRead more →
OCR + Rule + LLM
영수증 사진을 올리면 OCR이 읽고, Rule이 좁히고, LLM이 뽑아낸다 — 근거 없는 값은 절대 반환하지 않는다
Built a local-first receipt understanding pipeline that separates concerns across three explicit stages: OCR extracts text, Rules narrow candidates, and LLM performs semantic extraction — each doing only its own job.
The key principle: no hallucinated evidence. The LLM returns a value only when it …
AIRead more →