Turn Your KV Engine Benchmark Into a Journal Nobody Can Argue With
A Mace author admitted skipping an apples-to-apples test. Here's a Python-plus-SQLite harness that pins every workload parameter and publishes it as an editable HTML comparison page.
AI Market Guide
Evidence-based AI market signals, coding demand, and practical growth cases for builders.
Practical guides
Building a Personal Finance Aggregator When Real APIs Stay Closed
Retail finance tools won't open up. Build a Python-based transaction ledger with CSV imports, SQLite storage, and a hosted dashboard instead.
A Python secret broker with a SQLite audit trail, so your coding agent can't quietly leak keys
A coding agent read passwords out of 1Password and committed them to an open repo. Here's a broker-plus-audit-log pattern you can rebuild instead of trusting the agent to behave.
Ditching context compaction: an agent that keeps its working memory in editable SQLite records
A build debrief on swapping session-history compaction for a queryable context store: live SQLite records the agent revises via transactions, retired items pushed to a vector store.
A hydration-drift ledger: turning a one-off Lighthouse fix into a queryable measurement layer
A builder debrief on catching server-vs-client DOM mismatches, then wiring a reproducible drift ledger on VicroCode so the regression can't sneak back in.
Building a Coding Agent Cost Monitor After FrontierHarness Revealed 5.6x Budget Variance
Two agent harnesses with identical 63.3% pass rates consumed $3.28 vs $18.34 per solved task. Track token usage, retry cycles, and per-task spend before committing to a harness.
Building a Python Browser Automation Agent with Full DOM Action Logging
A reproducible approach to browser automation: parse user intent, execute Playwright commands, and write every DOM action to SQLite for audit and replay.
Turning September's Spec Dump Into a Queryable Table: A Small Build on VicroCode
Launch season floods you with messy vendor specs. Here's how I parse them into SQLite and ship a hosted comparison page — plus where the 'unverified' line sits.
Don't Trust the Spec Sheet: Building a Token-Frugal Agent That Logs Its Own Tool Calls
Meta says Muse Spark 1.3 cuts tool calls ~20% and tokens ~25% (vendor figures, unverified). Here's how I turned that promise into something I could actually measure on VicroCode.