CLOS
CLOS is a 37-agent cognitive operating system for modeling distributed cognition.
Studies: CLOS studies how many specialized agents can coordinate as a cognitive operating system without collapsing every task into one general model.Produces: It produces orchestration traces, agent role maps, and structured outputs that show how distributed cognition was assembled for a task.
Trinidad
Trinidad is an archival intelligence system over 22,000 Texas historical records from 1780 to 1900, deployed at trinidad.town.
Studies: Trinidad studies how historical records can become an evidence-aware retrieval and reasoning surface while keeping source material visible.Produces: It produces source-linked retrieval paths and archival intelligence outputs over the Texas historical corpus.
DRAMATURG
DRAMATURG is a compound instrument generating provenance-enforced screenplays from the Trinidad historical corpus.
Studies: DRAMATURG studies how generative writing can remain bound to a known archival corpus instead of drifting away from evidence.Produces: It produces screenplay drafts paired with provenance references back to the Trinidad historical corpus.
CORTEX
CORTEX is a 14-agent manufacturing intelligence platform for studying industrial AI workflows.
Studies: CORTEX studies how multi-agent intelligence can support manufacturing contexts where latency, review, constraints, and operational consequence matter.Produces: It produces agent-coordinated manufacturing intelligence workflows and reviewable reasoning traces.
EPPE
EPPE, the El Paso Proof Engine, is a civic blockchain notarization instrument for public accountability.
Studies: EPPE studies how public evidence can be notarized in a civic accountability context without asking users to trust an opaque record.Produces: It produces civic proof records, notarization events, and audit paths for public accountability workflows.
VERDICT
VERDICT is a 4-agent legal intelligence instrument with JUDGE, CLERK, WITNESS, and COUNSEL roles.
Studies: VERDICT studies how legal reasoning can be structured across explicit roles while preserving human judgment and evidence boundaries.Produces: It produces role-separated legal reasoning artifacts with traceable claims and review points.
MORTEM
MORTEM is a real-time biometric streaming and audit ledger instrument.
Studies: MORTEM studies how biometric signals can be streamed, interpreted, and audited without losing the chain between signal, event, and ledger entry.Produces: It produces biometric stream records and audit ledger entries for reviewable signal history.
AXON
AXON is an autonomous agent with Earned Autonomy Architecture, a behavioral backbone, and a blockchain audit trail.
Studies: AXON studies how an autonomous agent can earn autonomy through measured behavior rather than receiving unrestricted agency by default.Produces: It produces behavioral state records, autonomy classifications, and audit trail events.
GALEN
GALEN is a biomedical literature retrieval system built as a research instrument.
Studies: GALEN studies retrieval over biomedical literature where source grounding, narrowing, and careful evidence handling matter.Produces: It produces retrieved literature sets and source-grounded summaries for human review.
RECALL
RECALL is a semantic search RAG product for retrieval-augmented knowledge workflows.
Studies: RECALL studies how semantic search and retrieval-augmented generation can expose evidence paths instead of returning unsupported answers.Produces: It produces ranked semantic retrieval results, context bundles, and response traces.
Local-First AI Architecture
Local-first AI architecture keeps inference, routing, and evidence handling close to the data and the operator where possible.
Studies: privacy-sensitive, latency-sensitive, and context-sensitive AI systems.Produces: architecture patterns for local inference, evidence routing, and constrained cloud handoff.