Search
Lens searches configured organisational systems directly and, where enabled, optional local repository caches without exposing arbitrary destinations.
Digital Unity technology · Hexjack Lens
Hexjack Lens is a Secure-by-Design, read-oriented MCP server that gives AI systems controlled, evidence-backed access to an organisation's information estate.
Controlled research
Lens provides a 21-tool read-only MCP contract including research.investigate, with destination-bound source access, evidence provenance, runtime credential isolation and safe operational controls. Operational audit is metadata-only, runtime work is bounded, optional repository caches have an explicit retention model, and credentials rotate outside the application artifact. Cross-cutting security checks tie the public tool boundary, destination controls, secret safety, hostile-content handling, traversal protection and hardened runtime assumptions to executable evidence. Official releases carry immutable artifact identity, SBOMs, signatures and provenance, with verification paths for connected, restricted-network and native deployments. Lens remains standalone, with no Hexjack Agent Control dependency.
The information estate
Useful knowledge is distributed across documentation, issue trackers, source repositories, operational systems and other approved sources. Lens provides a controlled research layer between AI and that information.
Instead of teaching every AI client how every underlying system works, organisations configure their information topology once behind a bounded interface.
One Lens, bounded capabilities
Lens searches configured organisational systems directly and, where enabled, optional local repository caches without exposing arbitrary destinations.
Lens retrieves bounded source evidence while retaining source identity, context and provenance.
The public research.investigate operation delegates bounded questions through the provider-neutral Analysis Engine Contract. The configured provider can be local or remote and may own its own authorised workspace or analysis environment.
Evidence first
Lens is designed to preserve the distinction between recorded organisational evidence, observations established during authorised investigation and conclusions inferred from that evidence. Provenance tells an AI where information came from without turning every retrieved document into a trusted instruction.
Evidence & provenance →Secure by Design
Trust boundaries, least privilege, credential isolation, untrusted content, operational controls, cross-cutting security verification, software provenance and customer-verifiable release evidence are part of the Lens architecture rather than a hardening exercise at the end.
Explore Lens
Direct source retrieval, optional local repository caching, common evidence semantics and delegated investigation through the 21-tool read-only MCP surface.
Explore →Immutable software, external configuration and customer-controlled boundaries.
Explore →The engineering principles governing how Lens is designed, built and operated.
Explore →