Skip to main content
PromptGuard Enterprise provides multi-tenancy, SSO, role-based access control, persistent audit logs, IP allowlisting, and custom billing — all built-in. No separate deployment needed.For custom Enterprise plans, contact us at sales@promptguard.co.

Enterprise Features Overview

What’s Included

Not all of this is Enterprise-only. Organizations with RBAC, per-project roles and GDPR export and deletion are on every plan; SSO is also in the Shadow Team plan; and a Hybrid data plane (the engine on your servers, reporting to our dashboard) comes with Shadow Enterprise. Pricing is the full list.

Getting Started

1. Create Your Organization

/dashboard/* takes a dashboard sign-in, not an API key. These endpoints are the API the dashboard at app.promptguard.co calls. They are not part of the public API, and no API key or long-lived token works on them. Everything under /api/v1 takes your X-API-Key instead (API reference).To script one anyway, sign in with your email and password. This needs an account that has a password, and an email domain that requires SSO is refused. The access token lasts one hour.
Where an example shows YOUR_SESSION_TOKEN or YOUR_SESSION_COOKIE, it means the access_token from step 1, sent as Authorization: Bearer <token> or as the access_token cookie. A request acts on your personal organization unless it sends X-Organization-Id.

2. Invite Team Members

3. Set Organization Context

All dashboard API calls accept the X-Organization-Id header to scope operations to your organization:

RBAC (Role-Based Access Control)

Role Hierarchy

Enforcing Roles

Roles are enforced server-side. No client-side bypasses possible:

Self-Hosted Deployment

Run the entire PromptGuard engine, API, dashboard, database, cache, on your infrastructure. In the package’s default Air-gapped mode, prompts, verdicts and events stay on your network, with egress denied by your firewall; the Shadow AI agents on your fleet point at your engine instead of our cloud.
Self-hosting is part of the Enterprise agreement: you receive source access (Business Source License: you can audit exactly what the engine does with your data), a signed license file, and the deployment package described here. Contact sales@promptguard.co to get set up.

Deploy with Docker Compose

Prefer to watch? This section as a silent, step-by-step video: The deployment package ships a hardened Compose stack (API with read-only filesystem and dropped capabilities, Postgres, Redis, dashboard, and nginx as the single entry point to the API). The API and dashboard are signed, prebuilt images in PromptGuard’s private registry; your licence comes with a read-only pull credential and the release you are licensed for. From the package root:
The environment variables that matter most (deployment mode, ML inference, trusted proxies, licence) are in the self-host configuration reference.

Deploy with Helm (Kubernetes)

Prefer to watch? The Air-gapped install on Kubernetes as a silent, step-by-step video:
The database Secret carries three keys. DATABASE_APP_URL is the same server and database as promptguard_app, the role that row-level security serves requests as, and PROMPTGUARD_APP_DB_PASSWORD is the password the migration Job sets on that role. A Secret missing either leaves the pods in CreateContainerConfigError. Generate the four API secrets once (openssl rand -hex 32, or -hex 16 for the salt), keep them in your secret store, and pass the same values on every helm upgrade: changing the salt or the encryption key makes the provider credentials already stored unreadable, and changing the audit key makes the audit log’s existing seals unverifiable. The API refuses to start on a value that is short, repetitive or copied from a template. Optionally, and recommended, also set api.secrets.encryptionMasterKeys (Compose: ENCRYPTION_MASTER_KEYS) to 1: followed by openssl rand -base64 32. Stored credentials are then encrypted under a key derived for each tenant, and the master key can be rotated: add a new numbered entry (1:…,2:…) rather than replacing the old one, set api.secrets.encryptionActiveKeyVersion (required once there are two entries), restart, run python -m shared.services.secret_reencryption --target-version 2 in an API container to move every stored credential to the new key, and only then remove the old entry. With more than one API replica, keep the active version at the old number until every pod has restarted with the new entry; when introducing the first key, that number is 0. Once the job reports legacy_v0: 0, set ENCRYPTION_LEGACY_READS=false (Helm: api.env) so credentials stored before are no longer accepted under the old shared key. The chart includes HPA autoscaling, pod disruption budgets and ingress. It runs on a PostgreSQL and a Redis you operate, such as a managed cloud database or one run by a Kubernetes operator, and refuses to install until both Secrets and the four API secrets are named. On a fresh database the user in DATABASE_URL must be a superuser, or have CREATEROLE and BYPASSRLS, because the first migration creates database roles. The chart’s autoscaling, disruption budget, NetworkPolicy and alert-rule settings are in the self-host configuration reference, and how the licence is checked is under Licensing.

Point your Shadow AI fleet at your engine

The agent is identical in cloud and self-host modes, only the engine URL changes. Detection, redaction, and event storage all happen inside your perimeter.

Your own load balancer

If you put your own proxy or load balancer in front of the API (in place of, or ahead of, the bundled nginx), its read and send timeouts must outlast PromptGuard’s 570 s upstream timeout. A long completion sends nothing until it finishes, and a proxy that gives up first returns a bare 504 with no PromptGuard error and no logged event. In nginx:
Each proxy you add appends to X-Forwarded-For, so count it in TRUSTED_PROXY_HOPS (environment variables).

Compliance

Persistent Audit Logs

Administrative actions are recorded as audit events, each carrying an integrity_hash (SHA-256) in a hash chain for tamper detection. Querying interaction logs (the scans themselves):
Available filters:

GDPR Compliance

A copy of your data (Right to Access) is on request: email security@promptguard.co from the account’s address. There is no self-serve download yet: POST /dashboard/compliance/data-export answers with a job id and a message to contact support for the download link, not with the data itself. Data Deletion (Right to Erasure):
Cascading deletion of all user data. An audit record is created before deletion for compliance tracking. Delete your data lists what is and is not removed.

Monitoring

To scrape a self-hosted API’s /metrics, install the alert rules and build a dashboard, see Monitoring in the self-host configuration reference.

Integration Examples

Enterprise SSO Integration (OIDC)

On the hosted service, single sign-on is OpenID Connect only, configured per organization, and we set it up with you. It works with Okta, Microsoft Entra ID, Google Workspace, Auth0 and any provider that serves .well-known/openid-configuration.
SAML 2.0, the self-serve setup portal and Directory Sync (SCIM) are on self-hosted installs only. They are provided by WorkOS, which is not yet configured on the hosted service. See Single Sign-On (SSO) and Directory Sync (SCIM); if you need SAML on hosted PromptGuard, tell us.
Direct OIDC flow:
  1. User visits https://api.promptguard.co/dashboard/auth/sso/{org-slug}/authorize
  2. PromptGuard redirects to IdP with PKCE challenge
  3. User authenticates at IdP
  4. IdP redirects back with authorization code
  5. PromptGuard exchanges code for tokens, retrieves user info
  6. User is auto-provisioned (if enabled) and logged in
Configuration (stored in organizations.settings.sso_config):
Supported IdP Providers:

Support

Support is by email, at support@promptguard.co. Plans that include priority support, Enterprise among them, get the committed response time stated on the pricing page; on the others we answer as soon as we can, with no committed response time. No plan includes an uptime commitment or a faster response: either exists only if it is written into an agreement you sign with PromptGuard (Reliability & Status). Service status is at status.promptguard.co. For an Enterprise agreement, self-hosting or help with a rollout, contact sales@promptguard.co.

Next steps

Security overview

What each detector catches and how to configure it

Audit logs

The tamper-evident record of administrative actions

Self-host vs Cloud

What differs when you run the engine yourself

Managed service providers

Run PromptGuard for several client organizations
Last modified on November 10, 2026