Ingest
Multi-source evidence pipeline
Documents, transcripts, emails, and screenshots flow in from disparate reservoirs and emerge as a single searchable, cross-referenced corpus.
Showcase
What Stonewall does, how it works, and what it looks like when the system is running at operational tempo.
What it does
Stonewall is a corpus engine for institutional analysis. It ingests raw evidence, normalizes it into searchable form, links it across cases and characters, and surfaces patterns of behavior with chronology, citations, and provenance intact.
Analyze
Phenomenology pattern detection
A 274-pattern taxonomy classifies behaviors across the corpus, instantiated against artifacts so any pattern can be traced back to source evidence on demand.
Recall
Cast and chronology graph
Every cast member, case, date, and citation is indexed. Ask who, what, when, or where and the system answers from the underlying ground truth.
Operate
Tactical decision surface
Live deadlines, settlement posture, expert deadline status, and matter-level readiness — the system stays current with the work, not after it.
Architecture at a glance
Three layers cooperate: a prompt core that orchestrates behavior, a flat-file catalog that tracks every artifact, and a codex layer that enforces canon and chronology across the corpus.
Prompt Core
Stonewall Brain
The orchestration layer — defines retrieval priorities, operational guardrails, and the companion skills that handle specialized tasks.
Flat-File Database
Catalog System
The primary artifact index, searchable by character, pattern, date, case, or keyword. Derived indexes give alternative access paths into the same ground truth.
Codex Layer
Canonical References
Cast codex, case codex, artifact codex, phenomenology registry, and deposition codex govern identity, chronology, patterns, and matter posture.
Automation
Pipeline & Sync
Document ingestion, transcription, email consolidation, and bidirectional sync with the operational systems where the work actually happens.
Operational surfaces
The control plane spans the repository, the workspace, and external evidence reservoirs. These are the living surfaces, not historical snapshots.
Repository
Catalog indexes, codex references, source artifacts, and the analytics site. The flat-file database lives here, version-controlled and inspectable.
Workspace Integration
Matter registry, archive database, communication surface, and task backlog — with cross-database relations and morning operational briefs.
Evidence Reservoirs
Source documents, transcripts, exports, and screenshots flow from upstream reservoirs into the repository for normalization and analysis.
Public Portal
Aggregate dashboard with corpus search, deadline counts, pattern reach, and public-facing portfolio views. The operator console carries the authenticated operator record.