Alexandre Oliveira shares insights from over a decade of developing enterprise search solutions, culminating in the open-source release of Viglet Turing ES.
The platform integrates search engines like Solr and Elasticsearch with RAG pipelines and agent frameworks, designed to operate within secure networks.
Key design choices include running on existing infrastructure to avoid migration costs, ensuring auditable answers with source citations, and creating agents that interact with indexed content rather than just providing responses.
The tool supports hybrid retrieval methods combining keyword and vector searches, allows customization of models and embeddings, and offers TypeScript SDKs for frontend integration.
A live demo showcases the platform's capabilities without installation, emphasizing transparency, security, and flexibility for organizations requiring controlled data environments.
Original title: I spent 10 years building enterprise search for clients. Then I open-sourced all of it.
The AI system has determined that this news is clickbait/sensationalist: : The original title uses hyperbolic language ('spent 10 years') and creates curiosity about the open-source release, which is more sensational than informative. This has coincided with the opinion of the majority of users.