trenchantcyber.com

What It Is

Trenchant Cyber LLC provides on-premise AI and security platforms for regulated industries, built around the idea of “your weights, your kernel, your keys.” Its focus is not a typical cloud chatbot, but keeping models, vectors, prompts, and inference traffic inside the customer’s own servers and network perimeter, enabling private AI systems for industries such as legal, accounting, and manufacturing.

Core Capabilities and Data Privacy

The site repeatedly emphasizes 100% on-premise inference, 0 bytes egress, and customer-owned model weights. On the model side, it mentions tc-local-llm-v3 and open-weight foundation models, with domain-specific tuning, auditability, forking, and continuous operation to reduce vendor lock-in. Its Agent Runtime can be used for drafting, redaction, retrieval, and taking actions inside existing tools, with human confirmation and audit trails. On the security side, it also includes WARDEN DLP, model-aware L7 inspection, and a compliance posture aligned with HIPAA, SOC 2, and NIST.

Pricing and Trial

The main content does not disclose specific pricing, plans, licensing models, or hardware costs, and only offers a Book a demo option. There is also a new offline AI product, TC/SOVEREIGN, marked as available for pre-order on May 25, 2026. It focuses on no-internet operation, solar/wind power, preloaded models, and offline knowledge bases, but likewise does not provide actual pricing.

Pros and Cons

The main advantage is its clear positioning: it is suited to organizations that are highly sensitive to data egress, cloud-based training, vendor kill-switches, and compliance audits. Local GPU racks may also reduce cloud round-trip latency and improve data sovereignty. The downside is limited transparency: there is no pricing, deployment timeline, API/SDK information, integration list, Chinese-language capability details, or model benchmark data. Its “0 bytes egress” claim is also limited to AI inference traffic under the standard configuration; customers still need to manage other potential leakage paths themselves.

Who It’s For and Access from China

It is better suited to compliance teams in legal, finance, manufacturing, and healthcare-related sectors, especially enterprises that already have data centers, security teams, and budgets for private deployment. It is not ideal for small teams looking to quickly and cheaply try a SaaS AI tool. The main content does not specify access from China or supported payment methods, so these should be treated as “unknown.” For deployment in China, key issues to assess include hardware delivery, remote support, compliance adaptation, and local alternatives, such as privately deployed large-model platforms, local RAG/Agent platforms, or self-hosted open-source LLM solutions.