Cloud & Self-Hosting

Fabric AI is available as a managed cloud service or as a self-hosted application.

Cloud Hosted

The recommended way to use Fabric AI. Visit fabric.pro to get started.

• No infrastructure to manage
• Automatic updates and new features
• Enterprise-grade security and uptime
• Managed scaling and monitoring

Self-Hosted

Fabric AI is open source and can be deployed to your own infrastructure for full control over data and configuration.

Self-hosting requires:

• A cloud or on-premise environment capable of running containerized services
• A relational database
• Object storage for file uploads
• An AI provider API key

Detailed setup instructions, infrastructure requirements, and configuration guidance are available in the project repository.

Security

Self-hosted deployments follow the same security practices as the cloud version:

• All sensitive credentials are encrypted at rest
• Internal services are not exposed to the public internet
• Audit logging provides complete activity trails

Webhook IP Allowlists

If you enable webhook IP allowlisting for agent deployments, configure the proxy header Fabric should trust for the caller's source IP:

TRUSTED_PROXY_IP_HEADER=cf-connecting-ip

Supported values today:

• cf-connecting-ip
• x-forwarded-for
• x-real-ip

Choose the header your ingress or reverse proxy sets after normalizing the real client IP. Do not point this to a header that can pass through from the public internet unchanged.

Operational notes:

• Cloud-hosted Fabric manages this at the edge for you.
• Self-hosted Fabric should set this in the web application's runtime environment.
• If webhook IP allowlisting is enabled but TRUSTED_PROXY_IP_HEADER is missing, Fabric rejects those webhook requests by design.
• Keep webhook signature verification enabled even when using IP allowlists.

Next Steps