ARCHIVE
아카이브
기존에 공개한 글을 원래 주소 그대로 보관합니다.
과거의 글에는 작성 당시의 기술과 관점이 담겨 있습니다.
356개 글 · 4 / 12 페이지
· EN
Claude Opus 4.8 Dynamic Workflows
Claude Opus 4.8 Dynamic Workflows and Fast Mode: 1,000 parallel agents, real architecture trade-offs, cost data, known limitations, and an honest verdict on which teams should adopt this.
· EN
Ollama + FastAPI Local LLM Server — Docker Production Guide
Build a production LLM API with Ollama and FastAPI. Covers SSE streaming, health checks, Docker Compose. Llama 3.2 and Mistral execution logs included.
· EN
MCP Remote HTTP: Convert stdio Servers to Streamable HTTP
Convert MCP stdio servers to remote Streamable HTTP services. Covers stateful vs stateless session modes, Docker and Cloudflare Workers deployment patterns, with real experiment logs using TypeScript SDK v1.29.0.
· EN
Migrating from Jest to Vitest 4: A Complete 2026 Guide
Migrate Jest to Vitest 4.1.7 painlessly: config conversion, vi.importActual async traps, stable Browser Mode, and new matchers, all tested in a real sandbox.
· EN
Bun Shell Scripting Guide — TypeScript Automation Patterns
Bun Shell guide from Bun 1.3.14 experiments: $ template literals, .nothrow() error handling, Promise.all parallelism, macOS echo pitfall, and a comparison with zx.
· EN
Gemini API 2026 — Flash-Lite to Pro Speed & Cost Benchmarks
Real measurement data from May 2026. Compared Gemini 2.5 Flash-Lite (65 TPS), 2.5 Flash, 2.5 Pro, and 3.5 Flash under identical conditions.
· EN
Vector DB Comparison 2026: Qdrant vs ChromaDB vs pgvector Guide
Benchmarked Qdrant, ChromaDB, and pgvector on 1,000 vectors (dim=384): insert speed, query latency, and filter performance, with a decision guide for RAG.
· EN
I Tried RTK (Rust Token Killer)
RTK (Rust Token Killer) is a CLI proxy that compresses Bash output before it hits the LLM. I measured 90% token savings on find and 50% on ls, plus the limits.
· EN
Google I/O 2026 Antigravity 2.0
Antigravity 2.0 kills Gemini CLI on June 18. Teardown of the extension architecture, Gemini 3.5 Flash API, and what I/O 2026 means for the agent IDE war.
· EN
Next.js 16 + Claude API Streaming: Complete App Router Guide
Guide to Claude API streaming in Next.js 16 App Router: Route Handler, SSE handling, React 19 client hooks, error recovery — with real build code and logs.
· EN
AutoGen 0.7.x: Building Multi-Agent Systems from Scratch
Hands-on AutoGen 0.7.x guide. Build RoundRobinGroupChat, SelectorGroupChat, GraphFlow, FunctionTool step-by-step with real code, comparing with 0.2.x.
· EN
Claude Agent SDK Subagent Orchestration Tutorial
claude-agent-sdk 0.2.82 validated: AgentDefinition structure, parallel subagent spawning, and TaskBudget cost control. Full Python code walkthrough included.
· EN
Gemini 2.5 Flash Thinking API: thinking_budget Experiments
I ran Gemini 2.5 Flash's thinking_budget at 0, 1024, and 8000 across simple tasks, math reasoning, and code review. Simple tasks got 5x slower with no gain. Math reasoning actually reduced output tokens.
· EN
Claude Managed Agents' Dreaming, Outcomes, and Orchestration
A technical analysis of Dreaming, Outcomes, and Multiagent Orchestration announced at Code with Claude SF 2026.
· EN
Cloudflare Agents Week 2026: Agents Become Cloud Customers
Cloudflare Agents Week 2026: Sandboxes GA, Artifacts, Dynamic Workers, plus agents that open their own cloud accounts. Local @cloudflare/agents SDK notes.
· EN
AWS MCP Server GA: CloudWatch & IAM Tools for Claude Code
AWS MCP Server went GA May 2026. Connect 31 CloudWatch and 29 IAM tools to Claude Code. IAM condition-key architecture and honest real-world evaluation included.
· EN
Claude Agent SDK: Build Tool-Using AI Agents from Scratch
Build tool-using AI agents with Anthropic SDK. Covers tool definitions, multi-tool calls, error handling, and cost optimization in hands-on Python code.
· EN
FastMCP 3.x: Build a Python MCP Server in 30 Minutes
Ship a Python MCP server in 30 lines with FastMCP 3.2.4. The @mcp.tool(), @mcp.resource(), and @mcp.prompt() decorators expose tools Claude and Cursor call.
· EN
FastAPI + Claude API: Production Streaming API — SSE & Retry
A production FastAPI streaming guide with Anthropic SDK. SSE endpoints, exponential backoff retry, error classification, and Docker deployment covered.
· EN
Claude Code Masterclass #1 — Commands, Hooks & Subagents
Automate with Claude Code: slash commands for tasks, hooks for event triggers, subagents for delegation. A 3-step workflow guide using a real blog system.
· EN
Anthropic SDK vs OpenAI SDK: Developer Experience Compared
anthropic 0.100.0 vs openai 2.36.0, installed side by side: 408 vs 230 exported types, input_schema vs function.parameters, and which SDK really fits your work.
· EN
OpenAI Codex Cloud Agent vs Claude Code — May 2026 Comparison
OpenAI Codex became cloud-agent-only in April 2026. Architecture, benchmarks, cost, and workflow fit vs Claude Code — a clear May 2026 verdict for engineering teams.
· EN
uv for AI Development — Claude SDK Project in Under 1 Second
uv 0.11, the Rust-based Python package manager, installs 16 packages with anthropic in 0.874 seconds. Set up a Claude SDK project, pin versions, run CI/CD.
· EN
Gemini 2.5 Flash API Cost Optimization Guide
Gemini 2.5 Flash API: 4 cost-cutting techniques from direct experiments. Thinking tokens, Context Caching, Flash-Lite, Batch API. Real data proves 99% savings.
· EN
Anthropic Files API — Analyze PDFs Without Re-uploading
Upload documents once with Anthropic Files API and reuse across requests. Python SDK batch code, prompt-caching cost tradeoffs, and file_id management.
· EN
Google ADK vs LangGraph 2026 — I Installed and Compared Both
Google ADK vs LangGraph in 2026: sandbox comparison of code, deps, state management, branching, and CLI. Pick the right AI agent framework for your use case.
· EN
Langfuse v3 Self-Hosting — Complete LLM Tracing Guide
Run Langfuse v3 on your own infrastructure with Docker Compose: Python SDK 4.x instrumentation, RAG tracing, cost dashboards, and no cloud vendor lock-in.
· EN
Why Anthropic Cut Off OpenClaw
On April 4, 2026, Anthropic blocked Claude Pro/Max subscriptions from powering third-party agent tools. Add the Fast Mode 6x price multiplier, and this is a structural overhaul worth running the numbers on.
· EN
Anthropic's April Double Release
Claude Opus 4.7 (April 16) and Managed Agents beta (April 8) landed in the same month. Benchmarks are record-breaking but the community reaction is split.
· EN
PR Auto-Review with GitHub Actions + Claude Code CLI
Automate PR reviews with GitHub Actions and Claude Code CLI. Covers --bare, --dangerously-skip-permissions, and --max-budget-usd flags for safe CI execution, with complete YAML workflow examples.