lexset.ai

What It Is

Lexset mainly offers two product lines: Sentinel, an end-to-end perception system for field operations that detects objects from sensor video/data streams, performs cross-frame reasoning, geolocates findings, conducts QA, and generates reports; and Seahaven, a parametric 3D synthetic data platform for generating photorealistic, fully annotated training data. It is not a general-purpose chatbot, but an enterprise-grade visual AI system focused on defense, infrastructure inspection, spatial perception, and industrial robotics scenarios.

Core Capabilities

Sentinel’s key distinction is that it covers what happens after “detection” as well. For each detection, it attempts to generate latitude and longitude, using multi-level fallback methods such as telemetry, derivation, sensor inference, and frame-center estimation. It also includes working memory, tool use, cross-frame reasoning, and evidence citation, ultimately producing SALUTE reports, status summaries, or change reports. On the ingestion side, it supports MISB-compatible KLV and covers EO/IR, radar, and mapping platforms. Deployment options include cloud, on-premises, and edge devices, with Jetson-class hardware even mentioned. Seahaven can generate pixel-level annotations such as RGB, semantic segmentation, instance masks, depth, normals, and 3D pose, making it suitable for training where real-world data is scarce, long-tail defects exist, or new sensors are being introduced.

Pricing and Integration

The official website does not disclose plans, pricing, free quotas, or trials. It only offers “Request a briefing” and email contact, which clearly indicates a custom enterprise sales model. There is more information on integration: Sentinel can connect with Firestorm, CoT, GIS, BIM, and CMMS, while Seahaven can fit into PyTorch, TensorFlow, and existing training pipelines. However, no public API documentation, SDK, or usage-based pricing details were found.

Pros, Cons, and Limitations

Its strengths are a complete workflow, an emphasis on auditable evidence, support for offline/edge scenarios, and coverage of high-value verticals such as defense ISR and engineering inspection. The downside is limited transparency: there are no public model benchmarks, false-positive rates, latency test conditions, privacy details, or service SLAs. Chinese UI and Chinese report-generation capabilities are also not specified. For ordinary developers or small and midsize teams, onboarding and procurement may be relatively difficult.

Who It’s For and Access from China

Lexset is better suited to defense contractors, government projects, system integrators, utility inspection teams, and enterprise CV/ML teams. It is not ideal for individual users looking for a low-cost SaaS vision API. The official site does not provide clear information about access from China, and because its customer scenarios involve CMMC/ITAR and defense, cross-border procurement, data export, payments, and compliance may all involve uncertainty. Domestic alternatives in China may include local vendors focused on visual inspection, edge AI, and synthetic data.