Contexta / Blog
Contexta — writing

Technical writing from Contexta

Performance testing, Claude Code workflows, anti-forgery trust signals, and the engineering behind Site Check. No buzzwords. Evidence over claims.

All site-check· 4 reading-results· 3 ai-builders· 2 performance· 2 pricing· 2 running-tests· 2 troubleshooting· 2 ai· 1 baseline· 1 billing· 1 blocked· 1 bolt· 1 bot-protection· 1 certificates· 1 claude-code· 1 debugging· 1 dev-tools· 1 embed· 1 errors· 1 feature-ideas· 1 feedback· 1 getting-started· 1 jmeter· 1 load-testing· 1 lovable· 1 nfr· 1 p95· 1 product· 1 regressions· 1 v0· 1 vercel· 1

JMeter vs Site Check — when each tool is the right one

Site Check is for fast verification at three levels. JMeter is for journey-driven, multi-region, distributed load. Here's where the line between them actually sits.

running-testsjmetersite-checkload-testing

How AI-built sites typically score on Site Check — and what to do about it

Patterns we see across AI-built sites: what they almost always pass, what they almost always fail, and the three changes that move the score most.

ai-builderssite-checklovablevercelv0bolt

Top Claude Code Skills We Actually Use (April 2026)

An honest audit of every Claude Code skill in our install, cross-referenced against trending round-ups. Ten skills that matter, three that are worth trying, and one we skip on purpose.

claude-codeaidev-tools

How to file a feature request that ships — what we look for in the queue

Most feature requests we get are too vague to act on. Here's what makes one ship the same week, and what gets quietly archived.

feature-ideasfeedbackproduct

Performance certificates explained — what £29 and £49 actually unlock

We sell two paid certificates. Each one issues a different verifiable artefact. Here's exactly what you get for each price point and where the cert lives.

billingcertificatespricingembed

Top 5 errors Site Check throws and what each one really means

Walk blocked, timeout, certificate failure, redirect loop, and the dreaded "no pages fetched" — translated into plain English with what to do next.

troubleshootingerrorsdebugging

Comparing baselines — how to know if your latest deploy is actually faster

Single performance scores are useless without context. Here's how the baseline comparison view works and what to look at first when a release lands.

reading-resultsbaselineregressions

Setting NFR thresholds that match the traffic you actually get

Most NFRs we see are aspirational. Here's how to set thresholds that catch real regressions without crying wolf.

reading-resultsnfrperformance

Why your site got the BLOCKED verdict — and what to do about it

BLOCKED isn't BROKEN. Here's what it actually means, why it happens to enterprise sites, and the three options to get a real reading.

troubleshootingblockedbot-protection

Your Site Check report explained — P95, error rate, and slowest page

The five numbers on every Site Check report and what each one tells you about real users. Skip the technical jargon, read the actionable bit.

reading-resultsp95performance

Site Check L1 vs L2 vs L3 — which test do you actually need?

Three test levels, three different questions answered. A practical guide to which one fits your site, your stage, and your budget.

running-testssite-checkpricing

Getting started with Site Check — paste a URL, get two verdicts in 30 seconds

First time on Contexta? This is the 60-second version. What Site Check does, what the two scans tell you, and what to do with the result.

getting-startedsite-checkai-builders