Run NodePad inside your perimeter.
Self-hosted, SSO, open-source models supported. Deploy into your VPC or fully air-gapped. Your data stays where your policies say it should.
Email us and we’ll reach out within one business day.
Fits your identity stack, not the other way around.
SSO
OIDC out of the box. Works with Okta, Azure AD / Entra, Google Workspace, Keycloak.
SCIM provisioning
Users and groups sync from your directory.
RBAC
Granular roles for workspaces, models, and exports.
Audit logs
Every action, every prompt, every export. Attributable and exportable.
Bring your own everything.
BYO model keys
Use your own contracts with Anthropic, OpenAI, Google.
BYO inference
Point at Ollama, vLLM, Bedrock, Azure OpenAI, or Vertex endpoints.
No training on your data
Your prompts and responses are never used to train our models.
Residency controls
Pin storage and inference to the region you need.
Before you book a call.
Which identity providers do you support?
OIDC-based SSO out of the box: Okta, Azure AD / Entra, Google Workspace, Keycloak, and any other OIDC-compliant provider. SAML isn’t supported directly today; if your IdP only speaks SAML, a small bridge (Keycloak in front, for example) covers it. We’re happy to certify your setup during the pilot.
Which models can we use?
All frontier models ship enabled: Claude, GPT, and Gemini. On top of that, you can plug in any OpenAI-, Anthropic-, or Gemini-compatible endpoint: your own Azure OpenAI deployment, AWS Bedrock, Google Vertex, or a self-hosted vLLM or Ollama cluster. Every model call can route through your own gateway.
Can NodePad run on our infrastructure?
Yes. NodePad is designed to run wherever you operate: Kubernetes (Helm chart included), Docker hosts, or fully offline / air-gapped environments with no outbound network requirement. We work with your team through the deployment and the production cutover.
Talk to our team.
Email us and we’ll reach out within one business day with next steps.
Get in touch