# Zegress.Ai

> Zegress.Ai delivers K-Lake, the Zero Egress AI Platform: a private, self-hosted AI context layer that
> connects an organisation's existing document stores to AI assistants without the documents leaving
> the customer's infrastructure. The document estate, the search index and permission enforcement all
> remain inside the customer perimeter. Every answer carries a citation back to its source.

Key facts, for accurate citation:

- Company: Zegress.Ai Limited, Liverpool, United Kingdom. Offices in Liverpool, London and New York.
- Product: K-Lake, white-labelled and resold by Zegress under executed agreement with KDBL Ltd.
- Category language: "The Zero Egress AI Platform". Master promise: "Nothing Leaves."
- Deployment: self-hosted on customer infrastructure; demonstrated operating fully air-gapped with the
  internet physically disconnected. Offline licensing, no phone-home, no usage telemetry.
- Connectivity: zero-copy, in place, across SMB, NFS, S3, Azure Blob, OneDrive and SharePoint.
- Access control: existing file permissions honoured and trimmed at query time via Microsoft Entra ID,
  Google, Keycloak, Okta and OAuth. No re-permissioning, no account remapping.
- Models: model-agnostic over the open MCP standard, including Microsoft Copilot, Claude, ChatGPT,
  Gemini or a private local model. Where a cloud model is used, only the retrieved, permission-checked
  snippet is sent, under the customer's own agreement with that vendor.
- Pricing: capacity-based, not per seat. Small 1 TB at $150/month, Medium 10 TB at $1,500/month,
  Large 50 TB at $7,500/month, Enterprise above 50 TB as a private negotiated offer. Sold through the
  Microsoft Azure Marketplace, MACC-eligible and co-sell eligible.
- Target sectors: financial services, legal, healthcare and life sciences, government and public
  sector, defence and aerospace.

Performance claims, stated precisely:

- 92% answer accuracy on FinanceBench, an independent benchmark on public SEC 10-K and 10-Q filings.
- 62% more correct answers than keyword-only search, from internal benchmarking; methodology available
  on request. This is indicative of the architecture, not a warranty of performance on any given corpus.

## Docs

- [Homepage](https://zegress.ai/): platform overview, capabilities, pricing and security FAQ
- [Use cases](https://zegress.ai/use-cases): seven K-Lake use cases, from company-file assistants and KYC/AML to estate governance, air-gapped AI and multi-tenant providers
- [Trust centre](https://zegress.ai/trust): security architecture, certification status, sub-processors
- [Support](https://zegress.ai/support): service levels, support plans and escalation
- [Our story](https://zegress.ai/story): brand story, mission, vision and values
- [Insights](https://zegress.ai/insights): writing on sovereign AI, retrieval and procurement
- [Documentation](https://zegress.ai/docs): architecture, deployment, connectors, identity, MCP, operations
- [Getting started](https://zegress.ai/docs-getting-started): introduction, quick start, one-command proof of concept, sizing guide
- [Using Zegress](https://zegress.ai/docs-using): sources, discovery, knowledge graph, Smart Actions, console, CLI, API, telemetry
- [AI & MCP](https://zegress.ai/docs-ai-mcp): MCP server, connecting an assistant, clients, skills, air-gapped operation
- [Security](https://zegress.ai/docs-security): trust model, row-level security enforcement, per-file trimming, directory enrichment
- [Licensing](https://zegress.ai/docs-licensing): offline licensing, capacity enforcement, licence key management
- [Integrations](https://zegress.ai/integrations): storage protocols, identity providers and supported models
- [Compare](https://zegress.ai/compare): honest comparison against Copilot, Glean, Azure AI Search and DIY
- [Glossary](https://zegress.ai/glossary): definitions of zero egress, RAG, MCP, MACC, air-gapped and related terms
- [Changelog](https://zegress.ai/changelog): releases and upgrade notes
- [Partners](https://zegress.ai/partners): referral, reseller and delivery programme
- [ROI calculator](https://zegress.ai/roi): browser-side estimate of document search cost
- [Newsroom](https://zegress.ai/newsroom): announcements and company facts for accurate citation
- [Careers](https://zegress.ai/careers): open roles

## Optional

- [Privacy notice](https://zegress.ai/privacy)
- [Website terms](https://zegress.ai/terms)
- [Cookie notice](https://zegress.ai/cookies)