Introduction

Skirre is a fast, developer-friendly search engine that makes it easy to add powerful search to your application. Index your data, configure relevance, and start searching, all through a simple REST API.

What is Skirre?

Skirre is a standalone search engine designed for speed and simplicity. It provides full-text search, filtering, faceting, and optional hybrid semantic search, all behind a clean REST API secured with API keys.

Unlike embedded search libraries, Skirre runs as a service you deploy once and connect to from any application. Unlike heavyweight enterprise platforms, it's focused, fast, and easy to operate.

Key Features

  • Full-text search: BM25 ranking with configurable field weights and typo tolerance
  • Filtering & faceting: filter on string, number, and boolean fields with a SQL-like syntax
  • Hybrid search: combine keyword and semantic (embedding) scores for better relevance
  • Synonyms: define term equivalences so users find results regardless of phrasing
  • Analytics: track top queries and no-result queries to improve your search experience
  • API key management: issue admin and search-only keys with optional CORS origin restrictions
  • TypeScript SDK: a lightweight, typed client with automatic retries and error handling

How It Works

Skirre organizes data into indices. Each index holds a collection of JSON documents with a defined schema (searchable fields, filterable fields, sortable fields). You push documents into an index, configure settings like synonyms and typo tolerance, and then query via the search endpoint.

text
# Create an index
curl -X POST https://search.example.com/v1/indices \
  -H "Authorization: Bearer sk_search_live_..." \
  -H "Content-Type: application/json" \
  -d '{"uid": "products", "primary_key": "id"}'

# Index some documents
curl -X POST https://search.example.com/v1/indices/products/documents \
  -H "Authorization: Bearer sk_search_live_..." \
  -H "Content-Type: application/json" \
  -d '[{"id": "1", "title": "Running Shoes", "brand": "Nike", "price": 12000}]'

# Search
curl -X POST https://search.example.com/v1/indices/products/search \
  -H "Authorization: Bearer pk_search_live_..." \
  -H "Content-Type: application/json" \
  -d '{"q": "running shoes", "filter": "price < 15000"}'

Next Steps

  • Quickstart: index your first documents and run a search in minutes
  • Authentication: understand API key types and security
  • SDKs: use the TypeScript SDK for a typed, ergonomic experience
  • API Reference: explore every endpoint in detail