Capabilities

Local AI, mesh, fleet, and mission software in one operating layer.

EdgeLance is built for teams that need modern AI and sensor fusion without giving up control of devices, models, data, or mission continuity.

EdgeLance camera mesh contact tracking showing adaptive feeds, tracker IDs, tags, and mission handoff

Product In The Field

Actual EdgeLance screens across mission hardware.

Click through camera operations, command map, courier/fleet posture, and evidence review without shrinking every view into a grid.

Camera Operations

Adaptive camera wall, event clips, and mission handoff across field hardware.

Mission Surfaces

Each role gets the surface it needs.

EdgeLance is not a single screen. It is a set of connected mission views that preserve the same truth across operators, analysts, medics, fleet admins, and command.

Map

Common operating picture, assets, overlays, and current risk. Useful because every role shares the same map truth without needing the same panel layout.

Product Architecture

Six layers from sensor input to mission record.

Sense

6 sensor types fused at the node. Camera, acoustic, RF/SDR, face, NFC, wearable.

RF-DETR + YOLO detection pipelineAcoustic drone/gunshot/vehicle classifier28+ TPMS protocols for vehicle fingerprintingFace recognition + NFC badge IFF9 wearable health providersLicense plate recognition

Understand

35 open-weight models. On-device inference. Evidence-coupled outputs with SHA-256 frame hashes.

Gemma 4, Phi-4, Llama, Mistral, WhisperSAM 3.1 segmentation + FastVLMMission RAG with vector searchLoRA mission adapters (<1s hot-swap)Recommendation engine per node tierEvery AI output linked to source frame

Route

9 transport types. Auto-degrade from Starlink to LoRa to silent. Priority-aware store-forward.

Starlink, Direct-to-Cell, LTE, WiFi, BLELoRa/Meshtastic, MANET, Iridium SBDUSB-C Courier (zero RF)CRDT state sync across meshClassification-aware payload shapingBurn tombstones override delayed payloads

Act

10 role-specific views. Shared map. AI directives. Medic CASEVAC. Burst comms.

Tactical map with MIL-STD symbologyCamera wall with event-driven AIMission cell encrypted chat + AICASEVAC 9-line with blood type/allergiesBurst sync across degraded linksObserver access via QR with scoped permissions

Control

Classification MDM. 12 security gates. Hardware trust. Duress wipe. Signed audit trail.

TOP SECRET / SECRET / CUI enforcementSTIG continuous compliance per deviceDevice passport with provenance scoringDuress PIN: silent wipe + decoy + alertMission Burn: AES-256-GCM key destructionHMAC-SHA256 tamper-evident audit ledger

Review

AAR generation. Evidence chain. Decision timeline. Sealed export for command and legal.

Auto-generated after-action reportSource-linked evidence with custody hashOperator accept/reject on every AI outputMission replay from event databaseSealed export with signed proof chainBurn receipt: data destroyed, proof survives

Technical Depth

More than a dashboard around AI APIs.

The product value is in orchestration: choosing where AI runs, how evidence is preserved, how mesh data moves, and how users challenge or trust outputs.

35-model inference broker

Routes detection, reasoning, segmentation, speech, and embeddings across Apple Silicon, NVIDIA, and approved cloud by policy, classification, device tier, and thermal state. Continuous batching. Speculative decoding. LoRA adapter hot-swap in under 1 second.

Mission planning with AI fit

Operators select team, nodes, models, and documents. Recommendation engine scores 35 models against each node's memory, tier, and connectivity. Fit check validates budget. Launch gates block until all checks pass.

Acoustic multi-classifier

Drone motor harmonics (90-420 Hz with harmonic analysis), gunshot/blast (high-frequency + dB spike), vehicle engine (40-180 Hz), speech (200-4000 Hz VAD), footfall estimation. Ukraine-validated approach on consumer hardware.

9-transport mesh with link simulation

Operators select available transports during planning. System simulates payload mode (full/summary/coords/burst/silent), AI mode, evidence mode, and burn reach for each link profile. Auto-degrades at runtime.

Hardware trust interrogation

Provenance scoring for every device joining the mesh. Firmware analysis. Default credential detection. CVE assessment. Chinese-origin IoT sinkholing. Automatic quarantine or enrollment denial.

Cryptographic mission lifecycle

Mission key generated with 32 bytes of entropy. All AI outputs, sensor events, and operator actions encrypted under mission key. Key destruction renders all derivative intelligence unrecoverable. Signed proof chain survives for accountability.

Packaging

One suite, or focused modules.

EdgeLance can be sold as the full mission suite or licensed around the parts a customer needs first.

EdgeLance Mission

The operator, medic, command, analyst, and review workspace.

EdgeLance Mesh

A stand-alone mission-aware mesh and data movement layer.

EdgeLance Fleet

Device, software, MDM, courier, and model readiness management.

EdgeLance Connectors

Integrations for cameras, radios, TAK-style tools, cloud LLMs, base GPU servers, and partner systems.

Walk through the product, not a slide deck.

The best evaluation is a guided mission flow: plan, live map, medic view, camera evidence, AI directive, mesh health, fleet readiness, and review.

Request Demo