About SALT19

An independent applied AI systems lab.

SALT19 researches and builds governed cognitive systems and practical software for real-world work. EvoMind is the flagship cognitive-architecture research program; AeroClear and SALT19's applied products translate the same systems discipline into software people can use in operational environments.

Engineering posture: local-first where practical, human-gated where consequential, and verifiable where actions matter.

Governed cognitive systemsLocal-first AIAutonomous reasoningPersistent memoryPlanning + executionVerification + provenanceUAS flight intelligence Governed cognitive systemsLocal-first AIAutonomous reasoningPersistent memoryPlanning + executionVerification + provenanceUAS flight intelligence

What SALT19 does

Researches and builds governed intelligent systems that can support real operational work.

SALT19's work is organized around three connected pillars: cognitive systems, applied products, and research + validation. The common problem is not simply generating an answer; it is building software that can maintain context, determine what needs to happen, act through bounded mechanisms, verify outcomes, and preserve evidence.

Cognition

Cognitive systems

EvoMind research across reasoning, memory, planning, persistent state, semantic capability resolution, governed desktop action, verification, and experience-driven improvement.

Products

Applied software

AeroClear, SignalLoom, Patriot Claim Tracker, Recruiter.ai, UAS software, and other focused systems turn the same engineering doctrine into concrete user-facing workflows and operational tools.

Evidence

Research + validation

Technical publications, controlled benchmarks, runtime proofs, acceptance tests, and explicit claim boundaries connect architectural ideas to observable evidence.

Current systems

Flagship cognition research backed by practical software.

The current portfolio is intentionally not presented as a collection of unrelated applications. EvoMind is the central cognitive-systems research direction; applied products show how SALT19 turns systems engineering into working software.

EvoMind

EvoMind

SALT19's flagship governed cognitive architecture for persistent reasoning, memory, planning, capability discovery and composition, local desktop execution, verification, and controlled experience-driven improvement.

Persistent conversational and cognitive state
High-level goals decomposed into required outcomes and capabilities
Governed execution separated from planning authority
Verified outcomes, artifact identity, provenance, and bounded experience reuse
Cognitive architectureLocal-firstGoverned autonomy
AeroClear

AeroClear

A live UAS flight-intelligence platform for drone pilots. AeroClear brings mission-relevant weather, airspace, cooperative aircraft awareness, terrain context, aircraft operating limits, launch-window analysis, and reporting into one operational decision surface.

Mission and launch-condition awareness
Weather, GNSS/Kp, airspace, terrain, and nearby aircraft context
Aircraft-specific limits and decision support
Free and Pro product experience with explicit operational boundaries
UASFlight intelligenceOperational software
SignalLoom

SignalLoom

Local-first workflow automation aimed at turning repetitive computer work into structured, inspectable operational flows rather than opaque one-off automation.

Workflow structure and repeatability
Local-first execution where practical
Operational evidence and human-visible control
Workflow systemsLocal-firstGoverned automation
UAS Survey

UAS Survey Tool

Applied UAS and geospatial software developed from SALT19's operational roots, supporting mission planning, spatial analysis, quality assurance, terrain-aware workflows, and export-ready outputs.

Mission planning and data-quality workflows
Terrain and geospatial analysis
Operational outputs rather than visualization alone
UAS engineeringGeospatialOperational analysis
Patriot

Patriot Claim Tracker

Local-first Chrome extension for organizing a veteran's own VA.gov claims, messages, documents, benefits, payments, health information, forms, and redacted exports.

VA.gov-focused local browser workflow
Explicit privacy and sensitive-data disclosures
Published on Chrome Web Store
Chrome extensionVeteran workflowLocal-first
Recruiter.ai

Recruiter.ai

Resume-first Chrome extension for local candidate profiling, exact job-post matching, freshness signals, missing-skill guidance, and heuristic ghost-job risk.

Local resume and candidate-profile processing
Public job-post matching with direct application links
Official Chrome Web Store listing
Chrome extensionCareer workflowLocal-first
Technical lineage

Earlier and specialized systems remain part of the engineering record.

Arbor Analyzer, UAS Thermal Analysis Tool, and other domain-specific projects reflect SALT19's earlier work in vegetation analysis, thermal inspection, geospatial processing, and specialized technical workflows. They remain relevant to the company's applied-systems history without defining the current headline portfolio.

Engineering doctrine

Governance is an execution architecture, not a slogan.

SALT19's current systems philosophy treats cognition and action as separable responsibilities. The objective is capable software that remains inspectable, bounded, recoverable, and accountable when it moves from inference into real work.

How the doctrine is applied

01 Reason before action.
02 Authority before execution.
03 Verification before belief.
04 Evidence before capability claims.
05 Human control at consequential boundaries.

Local-first direction

Keep sensitive files, workflow context, and consequential decisions close to the user where practical—then make external action explicit, governed, and verifiable.

Research + engineering

Architecture claims and validation evidence are published as different things.

SALT19 develops EvoMind as both software and a research program. Public technical work covers cognitive architecture, verified semantic transitions, governed experience compounding, capability composition, memory integrity, planning, and autonomous systems. Validation pages separately document what has actually been exercised in the runtime.

Validation record

The validation record distinguishes controlled demonstrations from unrestricted generality. Current evidence includes governed cognitive composition, multi-capability planning, real application work products, verified outcome admission, artifact provenance, persistent session continuity, bounded experience reuse, and controlled transfer work.

Architecture hypotheses are not treated as completed capabilities.
Internal acceptance evidence is labeled as internal validation.
Benchmarks are used as failure-discovery tools, not as an AGI certificate.
Open authority and convergence gaps remain visible in the validation report.

Trajectory

Domain engineering converged into a broader governed-computation thesis.

SALT19 did not begin by replacing its applied-technology roots with an AI label. The current cognitive-systems direction grew out of building software for real operational constraints, where state, evidence, reliability, human authority, and recoverability matter.

2020

Applied technology and autonomous systems

SALT19 begins with practical engineering across UAS, geospatial analysis, automation, and technical workflows.

Build phase

Specialized operational software

Domain applications turn field and analytical requirements into repeatable software systems, including survey, thermal, vegetation, and workflow tooling.

Applied AI

Local-first workflow intelligence and governed automation

The focus expands from specialized analysis toward software that can preserve context, reason over workflows, and act under explicit operational boundaries.

Today

EvoMind + product-backed cognitive systems research

EvoMind is now the flagship research architecture, while AeroClear and other applied systems demonstrate SALT19's ability to ship operational software alongside deeper cognitive-systems work.

Founder

Independent systems research led by an applied builder.

Gabriel Allit, founder of SALT19
Gabriel Allit

Founder · AI Systems Engineer · Autonomous Systems Developer

Gabriel Allit founded SALT19 in 2020 and leads its work across cognitive systems engineering, autonomous agents, UAS engineering, applied software, and technical research. He created EvoMind as SALT19's persistent governed cognitive architecture and developed AeroClear as a user-facing UAS flight-intelligence product.

His research focuses on autonomous reasoning, memory integrity, planning systems, desktop intelligence, governed execution, adaptive learning, verification, and human-supervised autonomy.

Claim boundary

Research evidence is not an AGI certification.

EvoMind is an experimental AGI-oriented governed cognitive architecture under active validation. SALT19 does not present it as externally certified AGI, unrestricted general intelligence, or independently replicated scientific proof.

Company direction

Build intelligent systems that can reason and act without giving up authority, verification, provenance, privacy, or human control.

Research, applied AI, autonomous systems, or operational software.

For technical collaboration, product inquiries, EvoMind research, AeroClear, UAS systems, or governed AI work, contact SALT19.

Contact SALT19 →