Skip to main content
Cascade offers a self-hosting option that gives you complete control over your observability infrastructure. Deploy Cascade’s backend and dashboard in your own environment while maintaining full ownership of your AI agent execution data, traces, and logs.

​
Why self-host?

Self-hosting Cascade is designed for organizations with specific operational and security requirements.

Data sovereignty

Keep all execution data within your infrastructure and geographic region

Network isolation

Run entirely within your private network or VPC

Custom integrations

Connect to internal monitoring systems and private infrastructure

Cost control

Scale based on usage patterns and leverage existing commitments

Latency optimization

Deploy in the same region as your AI agents

Compliance

Meet regulatory requirements for data handling

​
Architecture overview

Cascade’s self-hosted deployment consists of three main components.

​
Backend services

The backend handles trace ingestion and storage.
ComponentPurposePort
OTLP ReceiverReceives OpenTelemetry traces from SDK clientsgRPC 4317
REST APIServes trace data to dashboard and query endpointsHTTP 8000
DatabaseStores span data, attributes, and events-

​
Dashboard

The web application provides trace visualization and analysis.
  • Static frontend: React application served via CDN or web server
  • API client: Connects directly to your backend REST API
  • Authentication: Integrates with your API key system

​
SDK integration

Your AI agents continue to use the Cascade SDK with minimal configuration changes. 
init_tracing(
    project="my_project",
    endpoint="https://cascade.your-company.com/v1/traces",  # Your self-hosted endpoint
    api_key="your-internal-api-key"
)
The SDK sends traces directly to your self-hosted backend via OTLP HTTP.

​
Data residency

When you self-host Cascade, all sensitive data remains in your infrastructure.
Data TypeStorage Location
LLM prompts and completionsYour backend database
Tool inputs and outputsYour backend database
Reasoning extractionsYour backend database
Trace metadata and attributesYour backend database
Span events and exceptionsYour backend database
API keys (hashed)Your backend configuration
Project configurationsYour backend configuration

​
Deployment options

Cascade provides flexible deployment options to match your infrastructure.
Deploy Cascade on Amazon Web Services using managed services.Recommended architecture:
  • Compute: ECS Fargate or EKS for containerized backend services
  • Database: RDS PostgreSQL for trace storage
  • Load balancing: Application Load Balancer for API traffic
  • Storage: S3 for dashboard static assets
  • Networking: Deploy in private VPC with security groups
Benefits:
  • Native integration with AWS CloudWatch
  • Easy scaling with Auto Scaling Groups
  • Compliance with AWS security standards
  • Regional deployment for data residency

​
Getting started with self-hosting

1

Contact the Cascade team

Reach out to the team to discuss your self-hosting requirements and receive deployment packages.
2

Review architecture requirements

Evaluate your infrastructure needs, database requirements, and networking setup.
3

Deploy backend services

Set up the OTLP receiver, REST API, and database in your environment.
4

Deploy dashboard

Configure and deploy the dashboard web application.
5

Configure SDK endpoint

Update your AI agents to point to your self-hosted Cascade endpoint.
6

Verify connectivity

Test trace ingestion and dashboard functionality.
Self-hosting requires a separate license. Contact hello@runcascade.com for pricing and deployment support.