Developer of an in-memory C++ search engine delivering instant, error-correcting search through a RESTful API.
The engine used an in-memory indexing technique that integrated inverted-index, trie, and forward-index structures to reduce search latency. A benchmark commissioned by the company and published in 2014 reported more than 400 times faster error-correcting search and 19.5 times lower power per query than Lucene on Android 4.3.
The native C++ implementation avoided Java virtual machine overhead on Android devices, where the company positioned its SDK against Java-based search libraries. Instant prefix search, fuzzy correction, geo search, and real-time updates were combined in a single engine.
Source: finance.yahoo.com
SRCH2 competed against the widely adopted open-source Lucene ecosystem and commercial platforms such as Endeca, which offered larger communities and existing enterprise integrations. Its engine was proprietary, with source code published on GitHub under the SRCH2 organization.
Distribution relied on trial-license-key downloads for Linux, MacOS, and Android rather than a self-serve cloud service. Named early adopters included CBS, Huawei, and HTC, and the company operated as a venture-backed startup throughout its life.
Source: techcrunch.com
SRCH2's flagship offering was an in-memory search engine built in C++ that delivered fast instant, type-ahead search with on-the-fly error correction. The engine was distributed as Linux and MacOS server downloads and as an Android SDK for embedding search directly on mobile devices.
The engine exposed a RESTful API for search, insert, update, and delete operations and connected to MySQL, SQLite, Oracle, SQL Server, and MongoDB data sources. Deployment options included licensed, platform, and hosted arrangements configured stand-alone or alongside existing search infrastructure.
Source: web.archive.org