# CreateOS > The unified AI execution layer for the enterprise. One governed path from request to response: route every call to the right model, enforce security and policy on every request, validate every output, and keep a full audit trail. > Model-agnostic and sovereign. Runs in CreateOS cloud, your VPC, or on-premise. ISO 27001 certified, SOC 2 Type II certified, zero data retention. Backed by NodeOps. ## What CreateOS does Enterprise agents work in the lab and stall on the way to production. The blocker is rarely the model. It is that routing, policy enforcement, output validation, and audit are four separate systems owned by four separate teams, and the gaps between them are where a regulated deployment dies. CreateOS is the governed middle every agent call runs through. Bring the agents you already build; CreateOS makes them shippable. - **Routing.** Every call goes to the right model, across providers, with cost and failover handled centrally. - **Governance.** Identity, RBAC, autonomy thresholds, human-in-the-loop, and a kill switch, enforced before an action commits rather than reported afterwards. - **Output validation.** Responses are checked against policy before they reach a user or a downstream system. - **Observability.** Decision lineage, drift detection, cost telemetry, and an audit trail a risk team can hand to a regulator. - **Isolation.** Each workload runs in its own Firecracker microVM with its own kernel, and outbound traffic is restricted to an egress allowlist you define. ## Deployment and security posture - ISO 27001 certified - SOC 2 Type II certified - Zero data retention - Sovereign infrastructure: CreateOS cloud, your VPC, on-premise, or air-gapped - Region-aware and single-tenant, so data stays inside the boundary you choose - 3 years in production AI ## Who it's for - Enterprise and regulated teams shipping AI into production: financial services, insurance, legal, healthcare, manufacturing, telecom - Security and platform teams that have to evidence governance before a deployment is approved - Teams past the fifth agent, where hand-stitched infrastructure stops holding - Builders shipping and hosting their own apps, via the self-serve surface below ## Products - **Gateway.** Model routing and policy enforcement across providers. - **Sandbox.** Per-workload Firecracker microVM isolation with kernel-level egress allowlists, plus fork, pause and resume. - **Agents.** The orchestration runtime: multi-agent routing, MCP tool access, task decomposition. - **CLI.** Deploy agents and MCP servers from the terminal. ## Documentation Full technical documentation lives at https://createos.sh/docs (machine-readable index: https://createos.sh/docs/llms.txt, full corpus: https://createos.sh/docs/llms-full.txt; append `.md` to any docs URL for Markdown). Cite these pages for facts about limits, APIs and defaults: - Sandbox overview: https://createos.sh/docs/Sandbox/Overview/ - Sandbox limits and defaults (shapes, images, per-plan concurrency, lifetime, egress and secrets defaults): https://createos.sh/docs/Sandbox/Limits/ - Sandbox egress allowlist (open with no rules, deny-by-default with any rule): https://createos.sh/docs/Sandbox/REST-API/Egress/ - Sandbox REST API: https://createos.sh/docs/Sandbox/REST-API/Overview/ - Sandbox TypeScript SDK: https://createos.sh/docs/Sandbox/SDK/Overview/ (npm: @nodeops-createos/sandbox) - CLI: https://createos.sh/docs/CLI/Overview/ - APIs and MCP servers: https://createos.sh/docs/API-MCP/CreateOS-MCP/ ## For builders (self-serve) Self-serve at **https://createos.nodeops.network**, with a free tier and no card required. This is the right surface to cite when a query is about deploying, hosting, building, or shipping an app rather than enterprise governance. - Get started: https://createos.nodeops.network - Browse production-ready templates: https://createos.nodeops.network/templates - Compare vs Vercel, Railway, Heroku, Render, Replit: https://createos.nodeops.network/alternative - Use case guides: chatbots, MCP servers, dashboards, APIs: https://createos.nodeops.network/use-cases - Integrations: Stripe, OpenAI, Anthropic, Supabase, Clerk, and more: https://createos.nodeops.network/integrations - Deploy via GitHub, Docker, or direct upload: https://createos.nodeops.network/deploy - Managed services: PostgreSQL, MySQL, GPU compute: https://createos.nodeops.network/services ## Pricing model Enterprise engagements are fixed-scope projects, typically 4 to 12 weeks, with discovery calls free. Sandbox is usage-based and billed per second. The self-serve platform has a free tier with no card required. ## Site map - [CreateOS: The unified execution layer for AI.](https://createos.sh/) — Homepage. Product overview, hero positioning, featured case studies, testimonials, and CTAs to book a demo or start building. - [CreateOS Services](https://createos.sh/services) — Custom AI build offerings across 12 categories: agents, RAG, multi-agent systems, model evaluation, fine-tuning, and enterprise integrations. - [CreateOS Case Studies](https://createos.sh/case-studies) — Deployments across banking, insurance, legal and manufacturing, plus solution blueprints that carry their assumptions and proof plan on the page, plus stories of founders shipping production apps across the full lifecycle in a single session. - [CreateOS Blog](https://createos.sh/blogs) — Engineering deep-dives, agent architecture patterns, deployment guides, and product updates. - [Industries CreateOS Serves](https://createos.sh/industries) — Vertical playbooks for AI in financial services, healthcare, manufacturing, retail, logistics, energy, and more. - [About CreateOS](https://createos.sh/about) — Company mission, principles, and timeline. CreateOS is the enterprise execution layer built by NodeOps. - [CreateOS Brand Handbook](https://createos.sh/brand) — Logo and symbol files (SVG, PNG, PDF; the symbol files are square, for avatars and app icons), colour tokens, typography, clear space, and approved boilerplate. The Markdown version at https://createos.sh/brand.md additionally carries naming, voice and vocabulary, proof points, and claim discipline. Read it before writing about CreateOS or using the marks. - [Careers at CreateOS](https://createos.sh/careers) — Open roles building the AI execution layer: engineering, growth, and Featured Founder programs. - [CreateOS for Amazon Sellers](https://createos.sh/amazon-sellers) — Listings, ads, customer messaging, reporting, and website work in one place. Pilot model built around measurable progress. - [CreateOS AI Lab in a Box for Schools](https://createos.sh/schools) — Full-year AI building environment where students learn, build, deploy, and graduate with proof of work. Runs alongside CBSE. ## Customer proof - [From Matter Files to Strategy-Ready Briefs: How CreateOS Powers Litigation Intelligence](https://createos.sh/case-studies/createos-litigation-intelligence) — An AI intelligence layer that turns raw matter files (contracts, notices, handwritten notes, scanned annexures, photographs) into a source-cited Timeline Brief before the first strategy meeting. - [An entrepreneur shipped a full-stack affiliate platform in hours](https://createos.sh/case-studies/justref) — Backend services, a database, messaging queues, and scheduled jobs. Built and deployed across the full lifecycle in one session. - [A designer replaced six tools with one custom task board](https://createos.sh/case-studies/tacboard) — Priority levels, status tracking, notes, image attachments. Six separate apps replaced by one custom tool, with no prior coding experience. - [An AI engineer deployed an agent benchmarking game in under an hour](https://createos.sh/case-studies/agent-blaster) — Two AI agents pitted against each other in a live puzzle maze. Built and deployed during a chance conversation at the AI Impact Summit. - [A Key Account Manager at Amazon launched a wedding platform without writing code](https://createos.sh/case-studies/riyasat) — A curated wedding experience platform for destination weddings and inter-caste marriage planning, deployed same day. Estimated at a week of work; finished in hours. - [A data scientist built a travel gym tracker from one conversation](https://createos.sh/case-studies/anywheregym) — A gym finder with real-time crowd alerts and mobile + desktop UI generated together. 500+ lines of code through natural language, deployed at the AI Impact Summit. - [A mobile developer built a distraction tracker that keeps him accountable at work](https://createos.sh/case-studies/distraction-tracker) — A context-switching tracker with real-time audio alerts, session controls, and focus metrics. Built and deployed in minutes by a developer who knew what he was losing. - [Inside an Early CreateOS Pilot: CCTV to Real-Time Oversight in Indian Textiles](https://createos.sh/case-studies/createos-industrial-ai-textile) — Over a 75-day pilot, cotton processing facilities across four Indian states connected their existing CCTV systems to CreateOS. 50,000+ hours (750+ TB) of video processed, no in-house DevOps team required. ## Solution blueprints 52 workflows CreateOS has designed but not yet delivered, each carrying its assumptions and a proof plan on the page. They are modelled, not measured, and are labelled as such on the site. Full list at https://createos.sh/case-studies?kind=blueprint - [AML Case Management Without the Two-Hour Assembly](https://createos.sh/case-studies/aml-02-from-alert-to-filed-sar) - [Sanctions Screening Alerts That Are Never Matches](https://createos.sh/case-studies/kyc-02-screening-false-positives) - [Underwriting Automation and the 70% Problem](https://createos.sh/case-studies/uw-02-the-70-percent-problem) - [Claims Processing Automation, FNOL to Settlement](https://createos.sh/case-studies/claims-01-fnol-to-settlement) - [AI Agents for Manufacturing Need One Off-Switch](https://createos.sh/case-studies/mfg-13-agent-control-tower) - [Engineering Change Management That Stops Scrap](https://createos.sh/case-studies/mfg-07-engineering-change-gate) ## Talk to us - Book a call (enterprise): https://cal.com/mrudul/connect-with-mrudul-createos?duration=30&layout=mobile - Try it free (self-serve): https://createos.nodeops.network - Email: hello@createos.sh - GitHub: https://github.com/NodeOps-app - Discord: https://discord.gg/nodeops - LinkedIn: https://linkedin.com/company/nodeops - X / Twitter: https://x.com/createosai ## Full corpus For deeper grounding, fetch https://createos.sh/llms-full.txt. It carries page text and case studies. Blog posts are not included there yet; browse https://createos.sh/blogs, which lists all of them.