September 2, 2026 · Product Update

AeroClear is expanding. Three new SALT19 systems are moving forward.

AeroClear now has a broader aviation-data picture through FAA SWIM integration. DroneMesh, TacWing-X and CloneCode are advancing across drone-data production, UAS operations and local AI software engineering.

AeroClear UAS flight-intelligence dashboard
AeroClear · live UAS flight intelligence and mission planning
DRONEMESH · LOCAL WORKSTATION
Mission processing
Outputs
Coming soon

DroneMesh

Drone-data production

Preview →
TACWING-X · OPERATOR DISPLAY
Platform state
PNT VALID
GNSS + INS
LINK 92%
BATTERY 81%
WORLD TRACKS 04
AUTHORITY OPERATOR
In development

TacWing-X

UAS operations

Preview →
CLONECODE · LOCAL AI WORKSPACE
src/
 core/
 ui/
tests/
README.md
pyproject.toml
def verify(change):
  tests = run_tests()
  if tests.green:
    return review(change)

# local model route
mode = "Agent"
LOCAL AI

Plan prepared.
RED evidence captured.
Implementation bounded.
Fresh verification pending.
Active validation

CloneCode

Local AI engineering

Preview →
Live now

AeroClear is becoming a broader UAS flight-intelligence platform.

AeroClear already brings weather, GNSS and space-weather context, terrain, aircraft limits, nearby cooperative crewed-aircraft awareness, launch-window analysis, mission geography and flight reporting into one responsive planning surface. FAA System Wide Information Management (SWIM) integration broadens the aviation information available alongside those mission-planning tools.

For a pilot, the value is a more complete operational picture in one place instead of stitching together disconnected sources before a flight.

AeroClear dashboard showing the UAS mission-planning interface
AeroClear is decision-support software. It does not replace official FAA sources, required authorizations, aircraft manuals, visual scanning, pilot judgment or Remote Pilot in Command responsibility.
Coming soon

DroneMesh turns raw drone imagery into usable geospatial deliverables — locally.

DroneMesh is a private, local-first photogrammetry workstation built around a familiar desktop workflow: bring in flight imagery, watch processing advance, review outputs, and export the finished project without requiring SALT19-hosted GPU processing or routine customer imagery storage.

The current application surface covers orthomosaics, DSM/DEM and DTM products, hillshade, LAZ/LAS/XYZ point clouds, OBJ mesh geometry, DXF/SHP contours, reports and offline handover packages. Processing progress is visible in the application, and a local Qwen assistant can help users interact with the project through Ollama.

DRONEMESH · LOCAL WORKSTATION
Mission processing
Outputs
WorkflowDrag/drop → local processing → validated outputs
PrivacyRoutine imagery stays on the user’s machine unless exported
DeliverablesOrthomosaic · elevation · point cloud · mesh · contours
AILocal assistant through Ollama
In development

TacWing-X brings platform state, localization and operator control into one Android surface.

TacWing-X is an Android operator display and edge-autonomy foundation built around a vendor-neutral runtime. Its current physical validation target is DJI Air 2S with DJI Smart Controller RM500, using DJI Mobile SDK V4. The wider architecture is designed around platform adapters so future integrations do not require rebuilding the operator interface.

The current HUD surface is designed to keep live video where supported, heading and attitude, speed and altitude, resilient PNT integrity, confidence and drift, platform position, battery and link quality, obstacle and camera state where available, local ONNX perception overlays, world-track counts and freshness watchdogs visible to the operator.

TACWING-X · OPERATOR DISPLAY
Platform state
PNT VALID
GNSS + INS
LINK 92%
BATTERY 81%
WORLD TRACKS 04
AUTHORITY OPERATOR
Broader MAVLink, ROS 2/DDS, Autel, Parrot, PX4, ArduPilot and other adapters are not presented as implemented until adapter and hardware proof exists. Current actuation policy is operator-only; AI perception remains advisory.
Active validation

CloneCode is a Windows-first, local-first AI coding workspace.

CloneCode combines a modern coding workspace with local model routing, repository-aware context, editor and LSP intelligence, Git review, persistent sessions, skills, memory, connectors and governed agent workflows. It works with Ollama, LM Studio and optional loopback OpenAI-compatible local endpoints.

The primary workbench exposes Ask, Plan, Agent and Sandbox Agent modes beside the editor and assistant. Its coding loop is designed around reviewable evidence: plan, RED proof, implementation, GREEN, independent review and fresh verification rather than silently treating a generated patch as finished work.

CLONECODE · LOCAL AI WORKSPACE
src/
 core/
 ui/
tests/
README.md
pyproject.toml
def verify(change):
  tests = run_tests()
  if tests.green:
    return review(change)

# local model route
mode = "Agent"
LOCAL AI

Plan prepared.
RED evidence captured.
Implementation bounded.
Fresh verification pending.
PlatformWindows 11 x64 release lane
ModelsOllama · LM Studio · local-compatible endpoints
WorkspaceEditor · LSP · Git · local AI · sessions
Agent modesAsk · Plan · Agent · Sandbox Agent
SALT19 portfolio

Different products. A connected systems direction.

AeroClear focuses on the mission environment before and around flight. DroneMesh turns aerial imagery into deliverables after collection. TacWing-X moves closer to the operational aircraft and robotic layer. CloneCode applies the same emphasis on local execution, explicit authority and verification to software engineering.

Agent Utility Grid supplies machine-accessible technical services, while EvoMind continues the underlying research into persistent reasoning, memory, planning, governed execution and verified outcomes. The result is a portfolio spanning cognition, infrastructure, field operations and production work rather than a collection of disconnected demos.

Status boundary: AeroClear is live. DroneMesh, TacWing-X and CloneCode remain development-stage systems; their public pages describe current implemented surfaces and the next validation/release milestones without presenting unfinished scope as released capability.