Sigmie — a modern Elasticsearch library for PHP

What is Sigmie?

Sigmie is an open-source PHP library for Elasticsearch and OpenSearch — built for developers who care about search relevance, not about wiring up low-level mappings and analyzers. It ships a single fluent API for keyword search, vector and hybrid retrieval, faceted filtering, aggregations, autocompletion, RAG-style document chunking, and Laravel Scout integration.

If you're tired of stitching together raw Elasticsearch JSON, hand-rolling tokenizers, or paying a hosted search-as-a-service for queries you could run yourself — Sigmie gives you the ergonomics of a hosted service with the cost profile of an open-source library you control. Indices are versioned with zero-downtime aliases, mappings come with sensible defaults, and semantic search is one method call away.

Who is Sigmie for?

PHP and Laravel teams building search for SaaS, e-commerce, content sites, internal tools, and AI agents. If you're already running Elasticsearch or OpenSearch (or thinking about it) and you want a library that gets out of your way, Sigmie is the shortest path from "we need search" to "search is shipped".

What makes Sigmie different

• Fluent, typed PHP API — no raw JSON, no untyped arrays, just methods that compose.
• Hybrid keyword + semantic search out of the box, with RRF rank fusion and MMR diversity.
• Predefined property types tuned for relevance — name, email, address, price, tag, and friends — so you don't reinvent analyzers per project.
• Vector embeddings via a built-in embeddings service (or bring your own), 384-dim by default.
• Versioned indices with aliases for zero-downtime reindex and safe mapping migrations.
• RAG and agent-ready — passage chunking, citation-friendly retrieval, and a Model Context Protocol server at /mcp for AI agents.
• First-class Laravel integration — Scout driver, Laravel AI agent tools, and a fluent search builder that feels native.
• OpenSearch supported — drop-in connection swap; no vendor lock-in.

Get started with the 5-minute Quick Start, browse the Introduction, or jump into Semantic Search if you already know your way around Elasticsearch.