Headless API integration

Hosted search and discovery for any custom storefront.

Scouty's REST and GraphQL APIs cover product search, semantic retrieval, image similarity, document search, and grounded AI retrieval. Frontend components for major frameworks. Merchant dashboard for non-engineers.

Request a free expert-led Search Audit Book a demo

Platform-specific pain

Where Headless API stores typically struggle.

Building search infrastructure from scratch is expensive and slow
Stitching multiple vendors for keyword, vector, image, and document is painful
Merchant teams still need a non-developer dashboard
Edge performance and indexing costs add up fast
Headless stacks lack good off-the-shelf search components

Integration approach

How Scouty installs and stays in sync.

REST + GraphQL endpoints for products, documents, images, and AI retrieval
Indexing API with batched and incremental updates
Webhook-driven catalog and document sync
Pre-built React/Vue components (and a universal widget)
Merchant dashboard alongside the API
API key environments, rate limits, and observability

Capabilities

What you get on this integration.

Search API

Single endpoint with keyword, semantic, hybrid, image, and document modes.

Indexing API

Push products, documents, and assets via REST or webhook.

Document upload

Direct PDF and document upload, OCR, page-level indexing.

Image search API

Search by image URL or upload; visually similar suggestions.

AI retrieval API

Grounded retrieval calls with citations; pluggable LLMs.

Webhooks & events

Catalog updates, search events, AI completions, and audit hooks.

FAQ

Common questions about Scouty for Headless API.

Which frameworks have prebuilt components?

React, Vue, and a universal JavaScript widget at launch. Other frameworks can use the API directly.

How is the API authenticated?

API keys per environment with scope control. Public keys for browser-side search calls; secret keys for indexing and admin endpoints.

How do I roll back a bad index?

Index versions are tracked. You can pin a version, swap atomically, or run an A/B test before promoting a new version.

What's the latency target?

Our target is sub-200ms p95 for keyword search and most semantic queries. AI retrieval latency depends on the LLM provider you use. We share actual latency numbers with design partners during onboarding.

Can I use my own LLM?

Yes. Bring your own model or use Scouty's. Generation costs can be capped, prepaid, or passed through.

Get started

Ready to put one search bar across your products, documents, and assets?

Start with a free expert-led Search Audit. A Scouty specialist will review your storefront and recommend a clear next step.

Request a free expert-led Search Audit Book a demo