Show HN: Ext-Infer – Native LLM Inference and Embeddings for PHP

Category: library

Tags: llm, php-extension, in-process-inference

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 6)

Ext-Infer brings local LLM inference directly into PHP via a native extension, using llama.cpp through Rust bindings to run GGUF models in-process. This eliminates the need for Python sidecars or remote APIs, offering low-latency chat, embeddings, and reasoning with a fluent PHP API. Its novel approach of integrating LLM capabilities as a PHP extension with thread-safe design and Metal acceleration makes it a unique tool for backend developers working on AI-powered applications.

Target audience: backend devs

Repository: https://infer.displace.tech · Rust · MIT · 1 stars

View on Hacker News