Skip to content

Generative AI & AI Agents

Knowledge Navigator

Private, Offline AI Search

University of East London, UKPrivate AIRetrieval-Augmented GenerationEnterprise Search

Architecture at a glance

Local, controlled environment · no documents sent to an external cloud serviceLocal indexingSearch & explorationLocal documentscontrolled accessIngestionlocal preparationEmbeddingson local hardwareVector indexsimilarity searchSearch interfacequery & exploreLocal APIlocal endpointsSemantic searchsimilar contentLocal storageapp dataContainerpackaged deploy
Swipe to see the full diagram →Everything runs inside a controlled local environment: documents are embedded and indexed locally, then searched through local APIs.

In brief

Knowledge Navigator is a Directrix Technologies project for University of East London, UK: Private, Offline AI Search. Summary: AI-assisted document search that runs in a controlled local environment, without sending documents to the cloud. Outcome: Built a private knowledge assistant: local semantic search and AI-supported document discovery that run entirely in a controlled computing environment.

Overview

Knowledge Navigator makes large document collections easier to navigate with AI-assisted search, while keeping processing within a controlled local environment.

The system uses local document indexing, embedding-based retrieval, and an AI-assisted query workflow to support information discovery without requiring every document to be sent to an external cloud service.

What we built

  • A local document ingestion and indexing workflow.
  • Embedding generation for meaning-based retrieval.
  • Semantic similarity search.
  • API services for search and retrieval.
  • Interactive interfaces for querying documents.
  • Visualization and monitoring components.
  • Containerized deployment support for controlled environments.

How we built it

We prepared documents and generated embeddings locally, then indexed the resulting vectors to retrieve semantically similar content.

Local API endpoints serve search and retrieval, interactive interfaces support querying and exploring information, and the application is packaged in containers for consistent deployment.

The architecture was designed around local processing and controlled data access.

Services this project shows

Project outcome

Built a private knowledge assistant: local semantic search and AI-supported document discovery that run entirely in a controlled computing environment.

Potential applications

  • Internal enterprise knowledge search.
  • Technical documentation assistants.
  • Research repositories.
  • Data-sensitive information retrieval.
  • Offline-capable knowledge tools.
Next case studyNusqa — Unani & Ayurveda Mobile Platform
CallWhatsApp