Sold by: Openbox AI LLC
Deployed on AWS
Self-hosted AI agent governance platform that enforces policy, screens content, and records auditable trails - deployed entirely inside your AWS account.
Overview
OpenBox - The AI Trust Platform for Enterprise
OpenBox is a self-hosted control plane for running AI agents under enforceable policy. It gives each agent a verifiable identity, evaluates every action against centrally managed policy, screens model inputs and outputs for unsafe or sensitive content, and records an auditable trail of what agents did and why.
Deployed entirely inside your AWS account and VPC, OpenBox ensures no data ever leaves your environment - making it ideal for organizations with strict compliance and data-residency requirements.
Key Capabilities
- Agent Identity and Attestation - Agents are registered with cryptographic identities and credentials. Unregistered or mis-credentialed agents are refused rather than silently allowed, enforcing a fail-closed posture by default.
- Policy Enforcement - Open Policy Agent (OPA) evaluates every agent action against versioned policy bundles you control. Define behavioral rules, guardrails, and authorization boundaries centrally.
- Inline Guardrails - Screens for PII, toxicity, and NSFW content in real time. Optional GPU-accelerated prompt and response inspection via LlamaFirewall defends against prompt injection and jailbreak attempts.
- Durable Orchestration - Temporal-backed workflows for attestation, governance, and observability ensure long-running agent work survives restarts and remains traceable.
- Identity and Multi-Tenancy - Keycloak-based SSO with a dedicated realm per organization supports enterprise identity requirements.
- Web Console and REST API - Operate, monitor, and manage your entire agent governance layer through a unified interface.
The Trust Lifecycle
OpenBox enforces trust across five continuous phases:
- Assess - Baseline risk across 14 parameters and assign Trust Tiers (1-4).
- Authorize - Apply guardrails, OPA policies, and behavioral rules before agents act.
- Monitor - Gain real-time runtime visibility into drift, cost, and latency.
- Verify - Validate goal alignment and generate tamper-proof Proof Certificates.
- Adapt - Leverage human-in-the-loop approvals and policy suggestions derived from observed patterns.
Who It Is For
Engineering and platform teams deploying AI agents in regulated or data-sensitive environments - financial services, healthcare, public sector, and enterprises with data-residency obligations - who cannot send prompts, outputs, or audit data to a third-party SaaS.
Problems It Solves
- Agents acting without enforceable guardrails or an audit trail
- Sensitive data leaking through prompts and model responses
- Prompt injection and jailbreak attempts reaching production models
- No verifiable record of which agent did what, under whose authority
- Compliance and residency requirements that rule out hosted AI tooling
Works With Your Existing Agent Stack
One SDK, no architectural changes. OpenBox integrates with the orchestration frameworks you already use, including Temporal, LangGraph, LangChain, CrewAI, Mastra, n8n, and more. Integration takes three steps: generate an API key, install the SDK, and configure governance rules.
AWS Deployment
Delivered as a signed Helm chart (OCI) with every container image pinned by digest and an SBOM published per release. An AWS CDK template provisions an EKS Auto Mode cluster and installs the full stack end to end. Bring your own VPC, subnets, Route 53 hosted zone, and ACM certificate - the template imports your existing networking rather than creating new infrastructure. Public endpoints are restricted to a customer-supplied IP allowlist enforced at the load balancer, or can be made VPC-internal only. Optional GPU node support is available for LlamaFirewall acceleration.
From Shift-Left to Shift-Everywhere
Most tools watch one point and miss the rest. OpenBox checks every stage and every layer - from development through execution and monitoring - so nothing slips by. Every agent action is traceable, every output is verifiable, and every decision is governable.
Highlights
- Enforceable Policy and Guardrails for Every Agent Action: OpenBox evaluates every agent action against centrally managed, versioned policy bundles using Open Policy Agent. Inline guardrails screen prompts and responses for PII, toxicity, and NSFW content. Optional GPU-accelerated inspection via LlamaFirewall defends against prompt injection and jailbreak attempts.
- Complete Audit Trail with Cryptographic Verifiability: Every agent action, decision, and cross-system interaction is recorded in a tamper-proof ledger with full lineage tracking. Proof Certificates verify goal alignment and provide cryptographic evidence of what each agent did and under whose authority. This gives compliance, security, and platform teams the auditable record they need for regulated environments in financial services, healthcare, and public sector.
- Fully Self-Hosted in Your AWS Account - No Data Leaves Your VPC: OpenBox deploys entirely inside your AWS account via a signed Helm chart and AWS CDK template that provisions an EKS Auto Mode cluster. All container images are pinned by digest with an SBOM published per release. Public endpoints are restricted to a customer-supplied IP allowlist or can be made VPC-internal only. Supports bring-your-own VPC, subnets, Route 53 hosted zone, and ACM certificate for full network control.
Details
Sold by
Categories
Delivery method
Supported services
Delivery option
Helm Chart for Kubernetes
Latest version
Operating system
Linux
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on the duration and terms of your contract with the vendor. This entitles you to a specified quantity of use for the contract duration. If you choose not to renew or replace your contract before it ends, access to these entitlements will expire.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/12 months |
|---|---|---|
Enterprise | Unlimited agents, GPU support | $250,000.00 |
Dimensions summary
This listing offers one plan, the Enterprise tier, sold as a contract through AWS Marketplace. Pricing is based on units rather than per-agent or usage counts. The plan covers an unlimited number of agents, so your cost does not change as you add more agents. It also includes GPU support. Because there is a single tier, there are no scaling steps or add-ons to choose between. You commit to the contract term and receive the governance capabilities bundled in this plan.
Top-of-mind questions for buyers
The unit is the contract entitlement for the Enterprise plan, not a per-agent or per-event count. You get coverage for an unlimited number of agents under this plan. The unit represents your subscription to the platform rather than a resource you count or scale.
No. The Enterprise plan covers unlimited agents, so adding agents does not change your cost. Pricing is based on units for the plan, not on agent counts or event volume. There are no overage charges triggered by growth in agents.
GPU support is included in the Enterprise plan at no separate line item. This lets the governance platform operate alongside workloads that use GPU resources. You do not select or pay for a separate GPU add-on under this single tier.
openbox.ai
Helpful?
Vendor refund policy
All sales are final. Due to the nature of this product a one-time enterprise deployment license delivered as a license key refunds are not available once the contract is executed and the license entitlement is issued.
For questions about your purchase, deployment issues, or exceptional circumstances, please contact the Openbox team at devops@openbox.ai prior to executing your contract.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Helm Chart for Kubernetes
Supported services: Learn more
- Amazon EKS
- Amazon EKS Anywhere
Helm chart
Helm charts are Kubernetes YAML manifests combined into a single package that can be installed on Kubernetes clusters. The containerized application is deployed on a cluster by running a single Helm install command to install the seller-provided Helm chart.
Version release notes
OpenBox 1.0.1
A maintenance release of the OpenBox Helm chart for Amazon EKS. Install it the same way as 1.0.0.
LICENSING
- The chart now verifies licences issued by OpenBox's production licensing service (https://billing.openbox.ai ). Use the licence bundle and tenant ID you received from OpenBox; no extra settings are needed.
FIXES AND IMPROVEMENTS
- Policy bundles now compile natively on arm64 (Graviton) nodes as well as amd64. In 1.0.0 the backend had to run on amd64 nodes for policies to compile.
- Platform updates across the console, API and governance engine from the OpenBox 1.x line.
- The core services run as an unprivileged user inside their containers.
UNCHANGED
- reCAPTCHA stays off by default; in-cluster policy storage (no S3 required); organization and administrator created automatically from global.seedOrg.adminEmail.
- Upgrading a 0.x install is refused: install 1.x fresh.
IMAGES All images are tagged 1.0.1 in the Marketplace registry (linux/amd64 + linux/arm64; guardrails-api, model-host-* and vllm-openai amd64 only). New content in 1.0.1: openbox-backend, openbox-fe, openbox-core, openbox-core-license-verify, keycloak. The others carry the same content as 1.0.0.
Additional details
Usage instructions
OpenBox 1.0.1 - Helm chart for Amazon EKS
PREREQUISITES
- EKS 1.27+, a default StorageClass, an ingress controller (nginx, traefik or istio), kubectl and Helm 3.8+.
- A DNS zone you control. OpenBox serves 4 hosts: openbox-eks, openbox-api, identity and openbox-core under YOUR_DOMAIN.
- Your OpenBox license bundle (license.jwt, report-signing.key) and tenant ID, issued by OpenBox (billing.openbox.ai).
- Admin email on your company domain (public email domains are rejected).
- Optional: one amd64 GPU node for the LlamaFirewall model (vLLM).
- Upgrading a 0.x install is refused: uninstall it, delete its namespace, then install.
-
LOG IN TO THE REGISTRY aws ecr get-login-password --region us-east-1 | helm registry login 709825985650.dkr.ecr.us-east-1.amazonaws.com --username AWS --password-stdin
-
LICENSE SECRET kubectl create namespace openbox kubectl -n openbox create secret generic openbox-license --from-file=token=./license.jwt --from-file=report-signing-key=./report-signing.key
-
PUBLIC URLS Save as public-urls.yaml (replace YOUR_DOMAIN): openbox-fe: extraEnv:
- {name: VITE_API_URL, value: "https://openbox-api.YOUR_DOMAIN"}
- {name: VITE_CORE_API_URL, value: "https://openbox-core.YOUR_DOMAIN"}
- {name: VITE_WS_URL, value: "https://openbox-api.YOUR_DOMAIN"}
- {name: VITE_KEYCLOAK_BASE_URL, value: "https://identity.YOUR_DOMAIN"}
- {name: VITE_KEYCLOAK_CLIENT_ID, value: "openbox"} openbox-backend: secret: stringData: FRONTEND_URL: "https://openbox-eks.YOUR_DOMAIN" ALLOWED_ORIGINS: "https://openbox-eks.YOUR_DOMAIN" KEYCLOAK_CLIENT_ROOT_URL: "https://openbox-eks.YOUR_DOMAIN/" KEYCLOAK_CLIENT_HOME_URL: "https://openbox-eks.YOUR_DOMAIN/" KEYCLOAK_CLIENT_REDIRECT_URIS: "https://openbox-eks.YOUR_DOMAIN/" KEYCLOAK_CLIENT_WEB_ORIGINS: "https://openbox-eks.YOUR_DOMAIN" KEYCLOAK_CLIENT_ADMIN_URL: "https://openbox-eks.YOUR_DOMAIN/" KEYCLOAK_CLIENT_POST_LOGOUT_REDIRECT_URIS: "https://openbox-eks.YOUR_DOMAIN/"
-
INSTALL helm install openbox oci://709825985650.dkr.ecr.us-east-1.amazonaws.com/openbox-ai/openbox/charts/openbox --version 1.0.1 -n openbox -f public-urls.yaml --set global.ingress.domain=YOUR_DOMAIN --set global.license.expectedTenantId=YOUR_TENANT_ID --set global.seedOrg.adminEmail=YOUR_ADMIN_EMAIL --timeout 20m No GPU node: add --set llamafirewall.enabled=false. Point the 4 DNS names at your ingress load balancer and add TLS (global.ingress.tls).
-
FIRST SIGN-IN kubectl -n openbox rollout status deploy/openbox-backend-service --timeout=10m The backend restarts once after install, then creates your organization and admin. Password: kubectl -n openbox get secret openbox-seed-admin -o jsonpath='{.data.password}' | base64 -d Open https://openbox-eks.YOUR_DOMAIN and sign in with the admin email. Change the password afterwards.
DEFAULTS YOU CAN CHANGE
- reCAPTCHA is off. To turn it on, create a Secret with OPENBOX_RECAPTCHA_SITE_KEY and OPENBOX_RECAPTCHA_SECRET_KEY, then set global.recaptcha.disabled=false and global.recaptcha.keycloakSecret=<secret>.
- Policy bundles are stored in the cluster (1 GiB volume, single backend replica). No S3 needed. To use S3 instead, set global.policyBundle.bucket, region and opaRoleArn. See values.yaml.
- Optional components, off by default: openbox-executor and openbox-evri-monitor.
TROUBLESHOOTING kubectl -n openbox logs deploy/openbox-backend-service -c keycloak-realm-bootstrap kubectl -n openbox logs deploy/openbox-backend-service -c register-organization
Support
Vendor support
For support with OpenBox, please contact the Openbox AI team directly.
Support resources include product documentation and the OpenBox GitHub repository for technical guidance, issue reporting, and community discussion.
For questions about deployment, configuration, licensing, troubleshooting, or refund requests, reach out to OpenBox AI through the contact options available on their website or directly via: devops@openbox.ai
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
Airia discovers shadow AI, enforces runtime guardrails across models and agents, and maps compliance evidence to EU AI Act, ISO 42001, and NIST AI RMF.
AlertAI provides a single-click Security & AI FinOps gateway for AI Apps and Agents with zero code changes. Enforce 9 Standards EU AI Act, NIST and GDPR compliance using 500+ pre-configured rules to block LLM, RAG, and MCP threats in real-time. Ensure enterprise-grade safety and budget oversight with automated OWASP, MITRE alignment and 6 months of free AI cost optimization.
VDF AI Enterprise Platform is a secure, private-cloud and on-premises AI platform for multi-agent orchestration, enterprise AI agents, private RAG, model routing, governance, and custom model workflows.
Credo AI helps enterprises scale AI responsibly with a centralized platform to measure, monitor, and manage AI risk.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.