Testing
LLMs hallucinate. They generate plausible-sounding text that is factually wrong, cite papers that don't exist, quote statistics that were never measured, and confidently describe events that never happened. If your product uses an LLM to generate content that users trust — customer support answers, medical information, legal summaries, financial
Testing
Authentication is one of the biggest time sinks in end-to-end test suites. If every test logs in before doing its actual work, you're spending a significant portion of your test runtime on a flow you've already validated hundreds of times. Playwright's storage state API
Testing
ML models degrade in ways that traditional software doesn't. Code doesn't spontaneously get worse — but a model trained on March data can silently degrade by June as the distribution of real-world inputs shifts. A retrained model can score higher on aggregate metrics while regressing on the
Testing
A/B testing AI features is harder than testing traditional software. Changing a button color produces a measurable, binary outcome — users click or they don't. Changing an LLM prompt produces different text outputs, and measuring which output is "better" requires defining what better even means for
Database Testing
Xata is a serverless database platform built on PostgreSQL with a TypeScript SDK, full-text search, branch-based development workflows, and an HTTP API. Its branch feature — where every branch gets an isolated database — makes it particularly interesting for testing: you can create a branch per pull request and get real database
Database Testing
PocketBase is an open-source backend in a single Go binary: embedded SQLite database, realtime subscriptions, file storage, auth, and an extensible Go framework — all in one 30MB executable. It's become popular for indie projects, prototypes, and small teams who want a real backend without the operational overhead. Testing
Email Testing
Resend has quickly become the developer-favorite transactional email API — clean SDK, React Email template support, and a generous free tier. But email is one of the most undertested parts of most applications. Password reset emails that don't send, welcome emails with broken links, and invoice emails with wrong
Security Testing
OpenFGA (Fine-Grained Authorization) is an open-source authorization system developed by Okta, based on Google Zanzibar. It lets you define relationship-based access control (ReBAC) models — "user X has role Y on resource Z" — and check them at runtime. It's the engine behind Auth0's FGA product
Database Testing
Nhost is an open-source Firebase alternative that combines PostgreSQL, Hasura GraphQL Engine, Auth (JWT-based), Storage, and serverless functions into a single deployable platform. It's popular with teams who want typed GraphQL APIs auto-generated from their database schema without writing resolvers. Testing Nhost applications involves several distinct concerns: GraphQL
Backend Testing
Medusa.js is an open-source headless commerce platform built on Node.js. Unlike Shopify or WooCommerce, Medusa is code-first: you own the backend, extend it with custom services and workflows, and connect any frontend. This power comes with responsibility — you need to test your customizations to ensure the checkout, order,
API Testing
LiveKit is an open-source infrastructure for real-time audio, video, and data — used by teams building video conferencing, live streaming, voice agents, and collaborative applications. Testing LiveKit-powered apps is notoriously tricky: WebRTC is browser-dependent, media streams are asynchronous, and real room behavior requires multiple connected participants. This guide shows you how
Database Testing
Backend-as-a-Service (BaaS) platforms have become a default choice for teams who want authentication, databases, storage, and real-time subscriptions without building backend infrastructure from scratch. Platforms like Appwrite, Nhost, PocketBase, Xata, and Supabase each offer Firebase-style convenience with different trade-offs. Testing BaaS-backed applications requires a different mental model than testing a