The Prototype

The Archive exists as a live system, not a paper proposal. Observation Packets are being ingested, archetypes are being proposed and refined, and Editions are being cut on a regular cadence. Everything described on the Noetic Archive page is implemented in the prototype at some level of maturity.

The prototype is intentionally narrow at the edges and deep at the core: a small set of modalities ingested rigorously, a strict release protocol, and an ontology that grows only when the evidence warrants it.

An automated Architect agent runs the refinement pipeline end-to-end. Newly ingested packets are screened against existing archetypes; those that don't confidently map enter an unmapped pool where convergent clusters surface as candidates for new archetypes.

Periodic deeper passes propose new groupings, sharpen boundaries between similar archetypes, and flag drift when an archetype's definition no longer matches its exemplars. Sensitive decisions — promotions, merges, splits — remain human-in-the-loop.

The Archive publishes frozen Editions of its taxonomy. Each Edition is a self-contained, Git-tracked package: canonical archetypes and packets in text-based, diffable formats, alongside generated databases, integrity checksums, and a public-facing scientific overview.

This is what makes the Archive reproducible: any claim it makes can be traced to the exact Edition — and the exact evidence — that produced it.

The prototype ingests EEG end-to-end today, and the Observation Packet schema is built to accept additional modalities — fMRI, fNIRS, MEG, ECoG, wearables, eye-tracking, full-body motion capture, and emerging BCI systems — without redesign.

Data flows in from two channels in parallel: Psyntient Ground, our dedicated contribution device, and third-party partners and preexisting datasets that meet the Archive's consent and compliance standards. External ingestion is already underway, ahead of Ground shipping.

The Archive's foundational unit pairs neural signal with first-person report — but the prototype also accepts unpaired neural data. Report-less packets reinforce existing patterns and surface as candidate clusters, while new archetypes are only ever defined where genuine phenomenological evidence is present.

Refinement runs in tiers — deterministic integrity checks, lightweight scans across the unmapped pool, and deeper periodic passes that handle ontology evolution. Budget and cadence are enforced structurally, not just by prompt, and a pre-release audit precedes every Edition cut.

We share architectural principles publicly; specific algorithms, prompts, and internal tooling stay inside the company.

The Archive is at the stage where the right collaborators meaningfully shape what it becomes. We are actively talking to:

• Researchers in neuroscience, phenomenology, contemplative science, and cognitive science with paired or unpaired datasets that could enter the Archive.
• Science Advisory Network applicants who want to contribute as independent consultants on specific projects shaping the ontology and release protocol.
• Device and dataset partners building neural, physiological, or behavioral instruments that belong in a multimodal archive of human experience.