Next.js + React
LiveUse @headband/react in App Router or Pages Router search surfaces.
npm i @headband/reactOpen recipe Integrations
One retrieval API, then a recipe for wherever your data lives: React, server-rendered apps, relational databases, commerce systems, and A/B/C search experiments.
Adoption path
Framework recipes
Live means a first-party package exists today. Recipe means the pattern is documented on stable REST/SDK primitives.
Use @headband/react in App Router or Pages Router search surfaces.
npm i @headband/reactOpen recipe Call the REST API from loaders/actions and hydrate client-side refinements.
fetch('/v1/search', { method: 'POST' })Open recipe Render static catalog shells and hydrate Headband search islands only where needed.
fetch('/v1/search', { method: 'POST' })Open recipe Use fetch against /v1/search from load functions or progressive-enhanced forms.
fetch('/v1/search', { method: 'POST' })Open recipe Index relational records with sync scripts, then render server-owned search pages.
curl https://headband.dev/v1/searchOpen recipe Sync connectors
Run a scheduled SQL export into /v1/bulk, then validate with /v1/search.
npm run sync:postgresExport MySQL rows through the mysql CLI and batch them into Headband.
npm run sync:mysqlMirror Shopify products/variants into a searchable catalog index.
npm run sync:shopifyMirror products/prices into an index for account portals, docs, and internal tools.
npm run sync:stripeBlueprint for a paid BataDB add-on: BataDB owns billing and product UX; Headband owns retrieval, indexing, and analytics.
npm run sync:batadbManaged scheduling, retries, logs, and secret storage belong in the connector runtime roadmap.
A/B/C testing
Use Headband indexes as experiment variants and split traffic in your app. Your app owns traffic allocation; Headband keeps retrieval outcomes beside query quality.
Split traffic across Meilisearch, Typesense, and Elasticsearch indexes for the same corpus.
Compare synonyms, typo tolerance, filters, and boosts without changing client code.
Compare lexical, vector, and hybrid retrieval before rolling a RAG flow to everyone.