Some documents cannot leave. We built for that first.
“Sovereign” is used loosely in this market, usually to mean a data-centre region on a map. At LexCore it means two specific, separately purchasable things — and this page is the detail, including where each one stops.
The sovereign tier
Your review runs on local models hosted inside our Nigerian deployment. The contract text is never sent to OpenAI, Anthropic or Google — no third-party model provider sees it. You buy it with credits like any other review type, and it is the cheapest line on the price list because local inference has no per-token cost to pass on.
Dedicated deployment
LexCore is installed in your VPC, your data centre or your rack. Documents, database, search index, models and audit trail all sit inside your network boundary, under your access control and your retention policy. Air-gapped installations are supported, and the licence is verified offline — there is no callback to us.
How the sovereign tier actually works
LexCore routes every model call on two axes: what kind of thinking it is, and which tier the matter is on. The tier decides which models are even constructed.
“Runs on local models” and “stays in your jurisdiction” are two different claims, and only the first one is about the model tier. Jurisdiction follows where the deployment is hosted. On the sovereign tier of our managed platform, that is our Nigerian deployment — so your text reaches no third-party model provider, and it sits on our infrastructure in Nigeria. If your requirement is that data never leaves your boundary, that is segment two, not this tier. We corrected this wording in our own product once already; we are not going to sell it back to you.
Three shapes of dedicated deployment
They differ in one respect that matters more than any other: whether anything leaves your network at inference time. Be clear with yourself about which one your policy actually requires, because the middle option is the one people buy by accident.
Local models only. No cloud model backend is ever constructed — not configured-and-unused, but absent, so there is no credential to leak and no route to misconfigure. This is the configuration for matters under a policy that forbids third-party processing outright.
The platform runs on your infrastructure, but reasoning calls go to your own frontier-model account — your Vertex or Azure OpenAI tenancy, your contracts, your DPA, your spend. Attractive because AI spend lands inside governance you already have.
No route out of the environment. Licence verification is offline, updates and the authority corpus arrive as signed media, and telemetry stays inside. Built for environments where the network boundary is the control.
Specialized multi-model sovereign intelligence
Rather than forcing a single generic model onto every legal task, LexCore deploys a complementary synergy of specialized open-weight models running entirely inside your boundary. Zero cloud telemetry, zero egress, and zero licensing lock-in.
Native multimodal visual comprehension for scanned Nigerian court processes, Certified True Copies (CTCs), exhibits, and complex tabular schedules — parsed locally with no external OCR sidecars.
Unrivaled structured JSON schema adherence and tool-calling precision. Guarantees 100% compliant contract redlines, playbook rule enforcement, and structured finding synthesis.
Reinforcement-learning-driven chain of thought. Step-by-step reasoning tokens verify statutory conflicts across the 546 Nigerian Federal Acts and record transparent audit trails.
8,192-token dense text embeddings powering instant semantic retrieval in PostgreSQL pgvector. Zero cloud cost, high throughput, and zero external dependency.
The licence does not phone home
An on-premise product that calls a vendor server on a timer is the first thing a bank’s security review finds, and rightly so — it contradicts the entire premise. Ours does not have that call to remove.
What we claim about sovereignty, and what we refuse to
The same line we hold on the security page, applied to this word specifically. Hold us to it.
Hardware-attested confidential compute. Running the platform in a confidential VM so you can cryptographically verify what is executing, rather than take our word and a DPA, is under active evaluation and is not a shipped control. If a remote-attestation requirement is in your policy today, tell us during scoping — it changes the timeline, and we would rather lose the deal than be vague about it.
How a dedicated engagement runs
Four stages. Pricing is agreed at the end of the first one, when both sides know what is actually being deployed.
Bring us your security review
The fastest way to scope a dedicated deployment is to send the questionnaire your risk team would send anyway. We will answer it, show you the platform on your own documents, and price the engagement.