On-box AI agents
Run AI agents with on-box LLMs and local GPU compute to assist authorized agency workflows.
Explore the layerBring your data together.
Put intelligence to work.
Keep control of every layer.
A secure platform for public safety and government.
A private data foundation for developers building offline.
AI agents. Security & SIEM. Analytics. All on-box.
Use Hawkeye to power your organization.
Or make it the foundation of your next application.
Connect operational records, case materials, and media inside an environment you control. Turn fragmented information into useful, protected intelligence for authorized teams.
Explore agency workflowsCAD, RMS, documents, and media in one private foundation.
On-box security/SIEM, tenant isolation, and agency access controls.
AI agents and analytics with local models and GPUs.
The Hawkeye Hypervisor is the bare-metal foundation beneath the private data lake. It runs directly on the physical server and lets developers deploy their applications on-box in isolated virtual machines alongside Hawkeye platform services.
Run AI agents with on-box LLMs and local GPU compute to assist authorized agency workflows.
Explore the layerMonitor security events, endpoints, and data access inside your Hawkeye deployment—even when air-gapped.
Explore the layerQuery and analyze private data locally. Explore patterns and prepare insights without sending data to an external analytics service.
Explore the layerChoose the footprint. Define the boundary.
Keep the architecture working for you.
Install the Hawkeye Hypervisor directly on your bare-metal server. Deploy your own application in an isolated virtual machine on the same box as Hawkeye services, with local data, databases, pipelines, and AI reached through scoped APIs and MCP.
Your data lake. Your PostgreSQL database.
Your local AI. Your application code stays your own.
Hawkeye provides the private data and AI services behind your application: a multi-tenant data lake, PostgreSQL database, ETL / ELT pipelines, structured and unstructured search, and on-box models and GPUs. Build your interface and business logic in a separate codebase, then connect through APIs and MCP Server. On bare metal, deploy your application directly on-box in an isolated virtual machine on the Hawkeye Hypervisor.
™PLATFORM SERVICESDeploy your application directly on-box in an isolated virtual machine alongside Hawkeye services.
Applications and platform services share the on-box infrastructure while retaining separate codebases and access boundaries.
01// Illustrative integration pattern
02// Your application remains a separate codebase.
03
04const context = {
05 tenant: "agency-workspace",
06 source: "authorized-case-materials",
07 services: [
08 "data-lake", "postgresql", "pipelines",
09 "structured-search", "unstructured-search",
10 "on-box-agents", "on-box-analytics"
11 ]
12};
13
14// 1. Authenticate with scoped permissions
15// 2. Request authorized data through an API
16// 3. Use on-box AI agents and analytics
17// 4. Return results to your application
Explore the services your application can use through Hawkeye, with operations kept inside your chosen deployment boundary.
Use Hawkeye as your application's private data foundation. Bring files, operational records, and media into the data lake, and use PostgreSQL for relational application data. Access the services through APIs and MCP while keeping your application in its own codebase.
Keep original files, documents, and media in the lake. Use the PostgreSQL database for structured records, relationships, and application state.
Ingest, validate, normalize, transform, and enrich information. Build repeatable pipelines from source systems to the datasets your application consumes.
Query structured tables and search unstructured documents, transcripts, and extracted media metadata. Return tenant-authorized results with links to their sources.
Configure isolation, replication, recovery, and GPU capabilities for the selected infrastructure. Validate each deployment against your application's operating requirements.
Vibe code with an on-box LLM, your approved datasets, and Hawkeye features. Plan, build, and test a separate application, then deploy it offline on your private foundation.
Preservation, provenance, traceability, and controlled access belong at the foundation of your data workflows.
Export supported forensic images in EWF (.E01) format. Verify image integrity and every segment, include hash manifests and chain-of-custody records, and require examiner review before release for court presentation.
Bring the evidence and its history together. Hawkeye's evidence workflow supports Expert Witness Format (EWF) export in .E01 for supported forensic disk and volume images, with validation and custody records for examiner review.
Associate the evidence with its case, tenant, acquisition record, and original integrity values. Preserve the acquired image with controlled access; use working copies for ETL, search, and AI analysis.
Illustrative outcomes. This page does not process or export evidence.
Deliver the .E01 image and all segments with an evidence inventory, SHA-256 manifest, verification report, acquisition and export notes, and chain-of-custody record.
Preserve video, documents, and third-party productions in their original formats with hashes and provenance. Export them as companion evidence when appropriate. A file bundle is not an original forensic disk acquisition.
Apply case permissions, legal holds, retention rules, and release approvals. Log exports in the on-box SIEM and keep AI summaries clearly identified as derived work products.
Prepared for court review, with the record to support it. Admissibility depends on lawful collection, validated methods, authentication, custody, and the applicable court rules; the .E01 format alone does not establish admissibility. Validate the configured workflow and receiving tools before operational use.
Make each step intentional—from intake to human review.
Bring case materials, RMS records, and witness statements into the case workspace.
Verify the case assignment and source context.
The Air-Gapped AI Data Factory™.