Articles for people who
ship the stack
Original rewrites on React, Node.js, TypeScript and AI — practical notes from the same engineering practice behind our operator software. Article bodies are in English.
Tagged: architecture
Beyond Linear Scripts: LangGraph Nodes, Edges, and Durable State
Replace brittle loops with graph workflows: specialized nodes, explicit state schemas, branching, human gates, and resilience paths.
1103 wordsRead articlePick RAG Frameworks by Workload, Not by LangChain Momentum
When search and PDF-QA are not agents, Haystack and LlamaIndex can beat a LangChain monolith on latency, deps, and debuggability.
1200 wordsRead articleArchitecting an AG-UI Frontend: Event Streams, Widget Registry, and Local LLMs
Stop asking models for JSX. Stream typed text/tool/widget events, validate props, register React widgets, and keep the rendering pipeline closed for modification.
2410 wordsRead articleHow to Choose a JavaScript Library for Enterprise Apps in 2026
A practical checklist for picking a JavaScript platform—requirements, components, data architecture, accessibility, tooling, testing, and prototypes—with Ext JS as one enterprise example.
1069 wordsRead articlePick an MCP stack for mixed Python and TypeScript teams
Standardize the expensive seam—protocol, framework, or full product lifecycle—so a second language does not spawn a second MCP delivery system.
1393 wordsRead articleResumable Islands on Bun: Design Lessons from the Stoneware Framework
How a server-first framework treats HTML as the default, scopes JavaScript to resumable islands, and handles static export, errors, CSP and deployment.
3283 wordsRead articleMeasure First: Why Early Optimization Makes Next.js Apps Harder to Run
See how premature memoization, broad client boundaries and stacked caches add complexity to Next.js apps, and how a measurement-first workflow keeps them fast.
2280 wordsRead articleBenchmark the Failure Path: Next.js, Remix and Stale Authorization Data
A case study of how a stale-permission bug reframes a Next.js versus Remix decision, and how to evaluate frameworks on mutations, cache lifetimes and debuggability.
2058 wordsRead articleWill Smarter Models Shrink the Agent Harness or Make It Matter More?
Why stronger LLMs may remove some orchestration code yet raise the stakes for permissions, evaluation and state, and how to decide which harness parts deserve investment.
1620 wordsRead article
About these articles
Request a 24h estimate
Need the same stack in a production operator layer? Send the brief — estimate within 24 hours.