The fastest way to add
AI search to your product
AI search to your product
Install UQA
Let Claude or Codex install it
Use https://uqa.cognica.io/ to install UQA and its usage skillsInstall directly
curl -fsSL https://releases.uqa-cloud.cognica.io/install.sh | shFree locally · No separate server to provision · Stored data stays off company servers
This applies to local UQA. Data you share with Claude or Codex is subject to each AI service’s policies.
Review the installer first
curl -fsSL https://releases.uqa-cloud.cognica.io/install.sh -o /tmp/uqa-install.sh
less /tmp/uqa-install.sh
sh /tmp/uqa-install.shThen tell Claude or Codex
Use UQA for the databaseAvailable locally today · UQA Cloud service is in preparation
Everyone gets stuck here
Build accurate search and recommendations with UQA
Word and semantic search surface the most relevant results, while relationships in your data can also power recommendations.
Pick a query and search 20 camping products.
Search examples for products, hotels, and documents
Technology already adopted by global search infrastructure
Public release notes from Lucene, MTEB, and Vespa document the adoption of UQA’s core search technology.

The open-source search technology behind Elasticsearch and OpenSearch.
UQA’s word and semantic search combination was added to the official codebase.
Apache Lucene
It evaluates AI search models on shared criteria and publishes results on Hugging Face.
UQA’s search model was registered as an official benchmark baseline.
MTEBAn AI search platform for operating search and recommendations at scale.
Search and ranking support for UQA’s search model was added to the official codebase.
Vespa