The system
One brain. Many senses.
Legacy AI is built as a small constellation of parts, each with one job and a clear boundary. This is the architecture we’re building: some pieces run today, others are in active development.
Cortex
The brain, it knows
The knowledge engine. It forges approved interviews into typed, cited Knowledge Cards, retrieves them with permission awareness, and refuses honestly when there’s no grounded source. Tenant isolation is enforced in the database itself.
Lens
The eye, it sees
The interface: a single, calm orb. Ask by voice or text; answers come back with their citations. Deliberately simple, so the product can evolve underneath without the surface changing.
Atlas
The map, it shows
The living map of what the company knows: sources, the forge pipeline, where each card came from, and what’s flowing right now. It makes the invisible knowledge base something you can actually see.
Thalamus
The relay, it governs
One place to tune every product parameter (models, modes, grounding, voice) defined once, resolved per part, audited on every change. Control that’s honest about what it can and can’t do.
Larynx
The voice, it hears
Speech in, speech out. Swappable transcription and synthesis so an interview feels like a conversation, not a form. Local or cloud, your choice.
Pharynx
The intake, it reads
Swallows the documents the company already has (PDFs, sheets, pages) and normalises them to clean text the brain can forge into knowledge.
Nexus
The mesh, it wires
The self-hosted layer it all runs on: it wires the systems and their stores together and deploys them on your infrastructure, in the EU. The ground the product stands on, not a feature of it.
The whole system exists to keep one promise: every answer is grounded and permission-aware, or it doesn’t come at all.