Send Half of Your Claude Code Traffic to a Local Model
How to split Claude Code between the hosted API and a local...
Developer tooling: GitHub Copilot vs Cursor vs Claude Code from 14 months of daily use, Mistral Forge for enterprise AI, and AI crawler behavior.
This page collects 15 articles, newest first. Editors, linters, and CI. The AI coding tools get compared from months of daily use rather than a weekend trial, and the verdicts aren't uniformly positive: some of them cost more in review time than they save in typing. Also covered are Biome against ESLint plus Prettier on the same monorepo, package manager benchmarks with real install times and disk usage, and the quality gates that catch what human review reliably misses. The CI pieces lean practical, down to the workflow YAML. Where a tool is genuinely better, that's what it says; where the free option wins, that's what it says too, which happens more often than the marketing suggests.
Everything else we've published is in the main archive.
Showing 1-12 of 15 articles
Send Half of Your Claude Code Traffic to a Local Model
How to split Claude Code between the hosted API and a local...
Which Local Model Is Best for TypeScript and React?
SWE-bench measures Python bug-fixing. Here is what actually matters when picking a...
Running Claude Code Fully Local Is Not Worth It Yet
A tested walkthrough for wiring Claude Code to free Docker Hub models,...
GLM 5.2 Review: We Tested the API and the Docs Lie
We called Z.ai asking for GLM-5.2 and got an error, not a...
Exit Cost: The Stack Criterion Nobody Ever Budgets For
Teams price how fast a library gets them shipping. Thirty years of...
ESLint 9 Flat Config Migration Guide for TypeScript
What extends does not carry into flat config, and the parser-without-plugin trap...
Vite 6 Environment API: A Guide to Multi-Target Builds
Vite 6's Environment API lets one config define SSR, edge, and client...
What's New in create-vite 9.1.2: The July 2026 Release
create-vite 9.1.2 fixes a crash when passing --eslint to non-React templates and...
Cognition SWE-1.7: Frontier Coding at $1.97 a Task
Cognition shipped SWE-1.7 on July 8, 2026: a coding model landing near...
Ornith-1.0: The Self-Scaffolding Open Coding Model
Ornith-1.0 is an open-source coding model family that learns its own RL...
Bun vs Node 24 vs Deno 2.5: 2026 Runtime Showdown Tested
Bun ships 1.2, Node 24 hit LTS, Deno 2.5 added npm parity....
Cursor AI for TypeScript: 20 Tips That Boost Output
Cursor AI is faster with TypeScript than with JavaScript. These 20 tips...