# NeueCode > NeueCode is a sovereign-AI company headquartered in Portland, Oregon, USA, with a regional office in Kuwait City. Its flagship product is **NeueCode 7 — the Govern AI Autonomous Enterprise System** (also called the NeueCode Platform): the on-premises autonomous AI engineering and AI-governance system for regulated organizations (banks, government, defense, energy, healthcare) whose source code cannot leave their network. "Govern AI" is NeueCode's descriptor for governing every AI action in the enterprise — its own agents and third-party coding assistants alike. Build with autonomous agents on local open-weight models on the customer's GPUs; govern every action — including external coding assistants like GitHub Copilot, Cursor, and Claude Code through the AI Governance Gateway (prompt inspection, redaction, blocking, signed audit) — and, in strict mode, sign offline-verifiable evidence that the agent's own code paths did not egress (app-level; a packet-level OS tier is optional). Tagline: "Build with AI. Govern every action. Keep everything sovereign." NeueCode also builds NeueTrace, an on-premise AI video intelligence platform for every environment (warehouses, offices, retail, airports, schools, government, healthcare, law enforcement). Everything is designed for on-premise, GPU-accelerated deployment; both products are fully bilingual (English/Arabic) and the go-to-market focus is Kuwait and the GCC. ## About - **Legal name:** NeueCode LLC - **Founded:** 2024 - **Founder & CEO:** Ayaad AlShamroukh - **Headquarters:** 5838 SW Beaverton Hillsdale Hwy, Portland, OR 97221, USA - **Regional office:** Kuwait City (Kuwait) — serving Kuwait and the wider GCC - **Languages:** English, Arabic - **Website:** https://neuecode.com - **Sales:** sales@neuecode.com - **Support:** support@neuecode.com ## Products ### NeueCode 7 — Govern AI Autonomous Enterprise System (the NeueCode Platform) - **URL:** https://neuecode.com/platform - **Category:** Sovereign AI engineering & governance (agentic AI / autonomous software engineering + AI governance for regulated organizations) - **Description:** The sovereign, on-premise AI coding agent for organizations whose data and source code cannot leave their infrastructure (banks, government, defense, energy, telecom, healthcare, government contractors & system integrators) — and, in strict mode, signs offline-verifiable evidence it stayed. - **Key capabilities:** - Sovereign coding agent on local open-weight models (Qwen3-Coder-class and others; drop-in loading, VRAM-aware multi-model gateway, OpenAI-compatible API — no per-token cloud bills); versioned model registry with eval-gated automatic rollback - Bounded ReAct agent loop: reads code, edits files as reviewable diffs, runs shell and tests, works with git, queries databases (Postgres, MySQL/MariaDB, MS SQL Server, Oracle, SQLite & MongoDB) — risky actions (write, shell, commit, DB change) pause for human approval; hard iteration/token/wall-clock budgets - Multi-agent orchestration: a supervisor delegates to specialist sub-agents (researcher, analyst, sysadmin, code-reviewer), each sandboxed to its own tools and budget; an optional autonomous goal mode is bounded, risky-tools-denied, and off by default - Grounded self-correction: a deterministic external verifier checks the agent's claims against what tools actually returned; untrusted tool, web, and memory content is fenced as data, not instructions - Three sovereignty modes, switchable live without restart: strict (the agent's cloud + network paths made structurally unreachable — an app-level control; a packet-level OS tier via nftables / loopback-only namespace is optional and off by default), hybrid (governed egress), open - Deny-by-default egress broker emitting cryptographically signed (RSA-PSS) per-run non-egress manifests; secret-scan blocks credentials before any socket opens - Client-verified, per-invocation RSA-PSS capability tokens scope every privileged machine-tool call; the client verifies them offline with only the box public key — the server never holds a key to the developer's machine - Agent Flight Recorder: tamper-evident, hash-chained record of every step, tool call, and approval (actions and byte-sizes only — never code, prompts, or reasoning) — verifiable by auditors offline, without trusting the vendor or operators - Compliance evidence export mapped to EU AI Act Articles 12, 13 & 14, DORA, NIS2, SOC 2 CC6 controls, ISO 42001, and SAMA/NCA (a mapping, not a certification — NeueCode provides technical controls and audit evidence that support the buyer's own governance and compliance program) - Enterprise identity: SSO via Active Directory / LDAP and SAML 2.0 for any IdP (Okta, Entra, ADFS, Ping, Keycloak, Google) — group-to-role mapping, passwords never stored; instant total deprovisioning across web, desktop, VS Code and API clients on the next request - Governance console: deny-by-default capability registry that ships locked down with a fail-closed floor; every class toggles org-wide and per role; least-privilege checks on every tool call - AI Governance Gateway: point third-party coding agents (GitHub Copilot, Cursor, Claude Code) at NeueCode as their model endpoint — every outbound prompt is scanned, classified, and policy-checked (allow / redact / block), then signed into the tamper-evident audit trail before any upstream call; the tools developers already use inherit the organization's egress policy and record - Break-glass dual control: the highest-risk actions can require a second admin's approval (a two-person rule enforced by the appliance); single-admin self-approval is recorded distinctly and raises an alert - Governed peer mesh: NeueCode instances register, trust, and test one another over MCP; strict mode blocks peer egress like any other outbound path - Compliance portal: a read-only auditor pass, a hardware-bound sovereignty certificate, and a signed, integrity-tiered evidence-export endpoint - MCP server: external agents (Claude Code, Codex, Cursor) can drive the sovereign agent over an on-prem MCP endpoint behind an admin-issued, revocable key (minimal tool surface today; richer verbs on the roadmap) - Optional cloud providers (OpenAI, Claude, GLM, custom): off by default, double-gated by both the sovereignty mode and each user's grant, with automatic local fallback; unreachable in strict mode - Six-type compounding memory with semantic recall, scoped per user and per project, plus reusable playbooks (versioned, org-scoped, approval-gated, off by default, plain-text only — never self-run code) - Clients: web workbench (bilingual EN/AR, RTL), native desktop app (Windows/macOS/Linux, client-side tool execution), VS Code extension, OpenAI-compatible API - **Deployment:** On-premise or fully air-gapped on customer GPUs; nothing phones home. One-command SSH install (hardware check → NVIDIA/CUDA → database → app → web tier). The backend ships as compiled native code — no readable source on the customer box. Licenses are per-customer and hardware-bound, with offline validation; the 7-day full-feature evaluation cleanly self-destructs unless licensed. - **Target customers:** Banks and financial services, government and defense, oil/gas/energy, healthcare, telecom, and government contractors & system integrators — regulated organizations in Kuwait, the GCC, and beyond - **Commercial entry point:** A scoped, full-feature on-prem pilot on one test repository (duration agreed up front), ending with a security review and exported audit evidence (https://neuecode.com/pilot). Resellers and system integrators can apply to the NeueCode Partner Program (https://neuecode.com/partners) ### NeueTrace — AI Video Intelligence Platform (every environment) - **URL:** https://neuecode.com/neuetrace - **Category:** Enterprise security software - **Description:** A complete enterprise video surveillance ecosystem with edge AI devices, cloud dashboard, Windows operations room, and device management UI. - **Key capabilities:** - Multi-camera AI processing (6-40+ cameras per device) - Two-tier face recognition (known persons and visitor tracking with deep-learning embeddings) - Real-time activity detection: fire, smoke, weapons, violence, vaping, cigarettes - People counting with directional entry/exit tracking - 7-day footfall heatmaps and gender-based analytics - Multi-branch unified cloud dashboard with 13 pages, real-time updates - Windows Security Operations Center desktop app with HUD-style interface - Auto-registration of cameras, locations, and devices - On-premise deployment with GPU acceleration - Resilient event-driven data pipeline with automatic retry and queuing - **Target customers:** Warehouses and logistics, offices and corporate campuses, retail chains, hospitality, education, government, healthcare, airports, law enforcement and public safety - **Deployment:** On-premise processing; cloud reporting of events is optional and can be disabled for air-gapped deployments (no video or biometric data is required to leave the customer network) ### Police.Live — Law Enforcement deployment of NeueTrace - **URL:** https://neuecode.com/policelive - **Category:** Law enforcement and public safety software - **Description:** Purpose-built AI video surveillance for law enforcement and public safety facilities — built on the NeueTrace platform (same engine, law-enforcement configuration). The police.live domain is being consolidated into neuecode.com. - **Key capabilities:** - Real-time threat detection: fire/smoke, violence, weapons, restricted substances - Facial recognition and identity management (on-device processing) - 24/7 AI camera monitoring (augments human surveillance) - People counting and crowd flow analytics - Real-time alert delivery with timestamped visual evidence - Works with any standard RTSP/IP camera (no rip-and-replace) - **Target customers:** Police departments (booking areas, holding cells, interview rooms), jails and detention centers, airports and border control, government facilities, security integrators ## Frequently Asked Questions ### What does NeueCode do? NeueCode builds sovereign, on-premise AI. Its flagship is the NeueCode Platform — a sovereign AI coding agent (autonomous software engineer, governance console, tamper-evident Agent Flight Recorder) that runs entirely inside the customer network on local open-weight models. NeueCode also builds NeueTrace — the AI video intelligence platform (weapon detection, face recognition, license-plate recognition, activity analytics) deployed across warehouses, offices, schools, government, and public safety. ### Who is the NeueCode coding agent for? Regulated organizations whose source code cannot be sent to a cloud AI service: banks, government and defense, oil/gas/energy, healthcare, telecom, and government contractors & system integrators — with a Kuwait-first, GCC-focused deployment and support model. ### Does source code ever leave the customer network? The platform runs on the customer's servers with local open-weight models. In strict mode, the agent's cloud and network paths are made structurally unreachable by a deny-by-default broker (an app-level control), and it signs an offline-verifiable per-run non-egress manifest; a packet-level OS tier (nftables / loopback-only namespace) is optional. In hybrid and open modes, the customer defines what is allowed and everything is recorded. ### Is NeueCode SOC 2 or ISO certified? No. Its compliance evidence export is mapped to frameworks including SOC 2 CC6 controls, ISO 42001, and the EU AI Act — a mapping, not a certification. NeueCode provides the evidence for the buyer's own compliance case. ### Where is NeueCode headquartered? Portland, Oregon, USA, with a regional office in Kuwait City serving the GCC. ### Are NeueCode products cloud-based or on-premise? The NeueCode Platform runs on-premise (or fully air-gapped) on customer GPUs. NeueTrace and Police.Live run on-premise with GPU acceleration; cloud reporting of events is optional and can be disabled for air-gapped deployments. ### Does NeueCode support Arabic? Yes. All products support English and Arabic with full RTL (right-to-left) layout — in the products themselves, not just the website. Customer support is available in both languages. ### How can I contact NeueCode? - Sales: sales@neuecode.com (demo and pilot requests: https://neuecode.com/contact) - Support: support@neuecode.com - Privacy: privacy@neuecode.com - Legal: legal@neuecode.com ## Technology Stack - AI inference: GPU-accelerated, on-premise, local open-weight models (Qwen3-Coder-class and others) - Architecture: Async microservices, event-driven pipelines - Face recognition: Deep learning embeddings, two-tier matching - Computer vision: Real-time threat and activity detection - Web stack: React 19, TypeScript, Vite, Tailwind CSS v4 - Backend: ASP.NET Core, Python AI services - Hosting: On-premise customer-controlled servers; AWS for the website ## Key Pages - [Home](https://neuecode.com/) - [Platform — the sovereign AI coding agent](https://neuecode.com/platform) - [Security architecture](https://neuecode.com/security) - [Kuwait & GCC](https://neuecode.com/gcc) - [Request a pilot](https://neuecode.com/pilot) - [Partner program — resellers & distributors](https://neuecode.com/partners) - [Cloud agents vs sovereign on-prem AI](https://neuecode.com/compare) - [Solutions: Government & Defense](https://neuecode.com/solutions/government) - [Solutions: Banking & Finance](https://neuecode.com/solutions/banking) - [Solutions: Oil, Gas & Energy](https://neuecode.com/solutions/energy) - [Solutions: Healthcare](https://neuecode.com/solutions/healthcare) - [Solutions: Telecom](https://neuecode.com/solutions/telecom) - [Solutions: Government Contractors & System Integrators](https://neuecode.com/solutions/integrators) - [Download the app](https://neuecode.com/download) - [NeueTrace V3](https://neuecode.com/neuetrace) - [About](https://neuecode.com/about) - [Contact / book a demo](https://neuecode.com/contact) - [Terms of Service](https://neuecode.com/terms) - [Privacy Policy](https://neuecode.com/privacy) - [SMS Messaging Program Disclosure](https://neuecode.com/sms) ## Comparisons (honest, public-source) NeueCode publishes balanced, public-source comparison pages — including where each competitor is the better fit. The generic category comparison (cloud coding agents vs sovereign on-prem AI) lives at https://neuecode.com/compare; named video-security comparisons below. Core differentiators: on-premise processing (cloud reporting optional, can be disabled for air-gapped sites), multi-threat detection on one platform, works with any standard IP camera, and no NDAA §889-prohibited components. - [NeueTrace V3 vs Verkada](https://neuecode.com/vs/neuetrace-vs-verkada) — on-premise GPU surveillance vs cloud-native VSaaS - [Police.Live vs Flock Safety](https://neuecode.com/vs/policelive-vs-flock-safety) — on-premise facility AI vs cloud LPR networks (no national data-sharing) - [Police.Live vs ZeroEyes](https://neuecode.com/vs/policelive-vs-zeroeyes) — broad on-premise multi-detection platform (gun detection is one of a dozen models) vs human-verified, DHS SAFETY Act-designated gun-detection specialist - [Police.Live vs iOmniscient](https://neuecode.com/vs/policelive-vs-iomniscient) — on-premise AI video analytics comparison - [Police.Live vs Scylla](https://neuecode.com/vs/policelive-vs-scylla) — on-premise AI threat detection comparison - [Police.Live vs Vaidio](https://neuecode.com/vs/policelive-vs-vaidio) — on-premise AI video analytics comparison ## Guides (informational) - [What Is a Sovereign AI Coding Agent?](https://neuecode.com/guides/what-is-a-sovereign-ai-coding-agent) — definition, how it differs from cloud copilots (Copilot, Cursor), on-prem/air-gapped deployment, and who needs it - [What Is Govern AI? Governing Every AI Action in the Enterprise](https://neuecode.com/guides/what-is-govern-ai) — definition of governing AI: every AI action (your own agents and assistants like Copilot, Cursor, Claude Code) under policy, human approval, and signed audit evidence; why access control alone is not governance; how NeueCode 7 packages it as the Govern AI Autonomous Enterprise System - [On-Premise & Self-Hosted AI Coding Tools, Compared](https://neuecode.com/guides/on-premise-ai-coding-tools-compared) — an honest look at the self-hosted, air-gap-capable coding tools (Tabnine, Refact, OpenHands, Tabby, Continue) and how to tell a real on-prem agent from marketing - [NeueCode vs Tabnine](https://neuecode.com/guides/neuecode-vs-tabnine) — honest comparison: both run on-premise, NeueCode adds offline-verifiable non-egress proof, on-prem AD/LDAP/SAML identity, and native Arabic for regulated GCC teams - [NeueCode vs GitLab Duo](https://neuecode.com/guides/neuecode-vs-gitlab-duo) — editor-agnostic sovereign coding (VS Code, Visual Studio, desktop, web) with provable non-egress vs GitLab-ecosystem agents - [NeueCode vs JetBrains AI / Junie](https://neuecode.com/guides/neuecode-vs-jetbrains-ai-junie) — one governed agent across editors plus audit-grade proof, vs JetBrains-IDE-bound agents - [NeueCode vs Qodo](https://neuecode.com/guides/neuecode-vs-qodo) — sovereign agentic coding with cryptographic proof of non-egress, vs a code-governance platform - [NeueCode vs Refact.ai](https://neuecode.com/guides/neuecode-vs-refact-ai) — a supported, sovereign product with audit-grade evidence vs self-maintained open source (main repo archived 2026) - [NeueCode vs OpenHands](https://neuecode.com/guides/neuecode-vs-openhands) — offline-verifiable non-egress proof beyond RBAC and customer-controlled data location - [NeueCode vs Mistral Vibe for Code](https://neuecode.com/guides/neuecode-vs-mistral-vibe) — model-flexible (open-weight, your choice) plus provable non-egress, vs a Mistral-model-centric stack - [NeueCode vs Poolside](https://neuecode.com/guides/neuecode-vs-poolside) — model choice, editor breadth, and audit-grade proof, vs a frontier-model-lab stack - [NeueCode vs VDF AI](https://neuecode.com/guides/neuecode-vs-vdf-ai) — sovereign software engineering with offline-verifiable non-egress proof and a governance gateway for Copilot/Cursor/Claude Code, vs a general-purpose on-prem agent platform (orchestration, model routing, cost control) - [NeueCode vs GitHub Copilot](https://neuecode.com/guides/neuecode-vs-github-copilot) — honest comparison: Copilot wins on GitHub-native depth, ecosystem, and IP indemnification; NeueCode runs local open-weight models on your own GPUs with signed, offline-verifiable non-egress proof — and its AI Governance Gateway can govern Copilot traffic itself - [NeueCode vs Cursor](https://neuecode.com/guides/neuecode-vs-cursor) — honest comparison: Cursor wins on speed and frontier cloud models; NeueCode runs local open-weight models on your own GPUs with signed, offline-verifiable non-egress proof — and its AI Governance Gateway can govern Cursor traffic itself - [NeueCode vs Claude Code](https://neuecode.com/guides/neuecode-vs-claude-code) — honest comparison: Claude Code wins on agentic depth, frontier models, and the MCP/SDK ecosystem; NeueCode runs local open-weight models on your own GPUs with signed, offline-verifiable non-egress proof — and its AI Governance Gateway can govern Claude Code traffic itself - [NeueCode vs PUMAS](https://neuecode.com/guides/neuecode-vs-pumas) — honest comparison: PUMAS is a runtime enforcement layer (discovery, monitoring, policy, audit) over agents built elsewhere; NeueCode 7 builds the engineering agents, serves local open-weight models on your GPUs, and governs every action as one sovereign per-server system — air-gap ready, with offline-verifiable non-egress proof - [Cloud vs On-Premise AI Video Surveillance](https://neuecode.com/guides/cloud-vs-on-premise-video-surveillance) — data residency, cost, security, and when each fits - [What Is Air-Gapped Video Surveillance?](https://neuecode.com/guides/air-gapped-video-surveillance) — definition, who needs it, how Police.Live runs with no internet link - [PDPL & SDAIA Surveillance Data Localization (Saudi Arabia)](https://neuecode.com/guides/pdpl-sdaia-surveillance-data-localization) — why on-premise is the safe default for KSA - [Is Flock Safety a Privacy Risk?](https://neuecode.com/guides/flock-safety-privacy-concerns) — factual summary of the reporting + the on-premise alternative - [ZeroEyes Cost & Pricing Guide (2026)](https://neuecode.com/guides/zeroeyes-cost-pricing-guide) — every verifiable ZeroEyes price figure with sources (vendor's own <$50/camera/month claim; public contracts ≈$168–$592/camera/year via StateScoop public records), what drives the price, where ZeroEyes wins, and how the subscription compares structurally to owning an on-premise platform