> ## Documentation Index > Fetch the complete documentation index at: https://docs.asymptotelabs.ai/llms.txt > Use this file to discover all available pages before exploring further. # Deployment Options > Choose how to run self-hosted Asymptote, and what your cloud needs to provide. > [!NOTE] > > **Planned release.** Self-hosted Asymptote is in development and offered to design partners through a Private Deployment engagement; the Helm chart and Terraform modules described in this section are not generally available yet. This section describes the planned release: chart names, values keys and paths may change before it ships. Today's Asymptote Cloud dashboard uses hosted services for sign-in and product analytics; the self-hosted release replaces them, as described in [System Architecture](https://docs.asymptotelabs.ai/self-hosting/system-architecture.md#hosted-services-in-asymptote-cloud). [Contact us](https://asymptotelabs.ai/contact) to take part. ## Quick decision * **No infrastructure to run?** Use [Asymptote Cloud](https://docs.asymptotelabs.ai/deployment/asymptote-cloud.md). * **Telemetry must land in your account first, but a hosted dashboard is fine?** Use [BYOC storage](https://docs.asymptotelabs.ai/deployment/byoc-storage). * **No third party may hold or read your agent telemetry?** Self-host, on Kubernetes, with one of the two setup paths below. ## Self-hosted: Kubernetes Asymptote runs on Kubernetes only, installed from one Helm chart. The chart is the same on every cloud. ### Two setup paths | Path | Use it when | Guide | | - | - | - | | **Reference Terraform, then the chart** | You are starting from an empty AWS account, Google Cloud project or Azure subscription | [Terraform](https://docs.asymptotelabs.ai/self-hosting/terraform.md) → [Kubernetes Deployment](https://docs.asymptotelabs.ai/self-hosting/kubernetes.md) | | **Your cluster, then the chart** | You already run Kubernetes and the managed services, and provision them your own way | [Kubernetes Deployment](https://docs.asymptotelabs.ai/self-hosting/kubernetes.md) | ### Requirements | Requirement | Details | | - | - | | **Kubernetes** | A version the release is tested on (listed in its release notes), with an ingress controller and a storage class for persistent volumes | | **PostgreSQL** | PostgreSQL 15 or later, managed by your cloud | | **Object storage** | A dedicated bucket with new-object notifications | | **Identity provider** | Any SAML 2.0 or OpenID Connect provider. See [SSO](https://docs.asymptotelabs.ai/self-hosting/sso.md) | | **LLM endpoint** | Amazon Bedrock, Vertex AI, Azure OpenAI, or any OpenAI-compatible API | | **DNS and TLS** | Host names and certificates for the dashboard, ingest and policy endpoints | | **From Asymptote** | Registry credentials and a license key | ### Sizing Size depends on how many endpoints you have, how much agent activity they record, and how long you keep it. Event volume drives ClickHouse's CPU, memory and disk; the other services are small and scale horizontally. | What drives size | Where it shows | | - | - | | Number of endpoints | Ingest and policy service replicas | | Events per endpoint per day | ClickHouse CPU and memory | | Retention period | ClickHouse disk and object storage | | Policy enforcement on or off | Policy service replicas and LLM endpoint throughput | Asymptote sizes your deployment during the engagement from your fleet size, a sample of your endpoints' event volume and your retention period, and sets the chart's resource values to match. ## External dependencies ### By cloud The services each release is tested with, and what the reference Terraform creates: | Requirement | AWS | Google Cloud | Azure | | - | - | - | - | | Kubernetes | Amazon EKS | Google Kubernetes Engine | Azure Kubernetes Service | | PostgreSQL | Amazon RDS or Aurora for PostgreSQL | Cloud SQL for PostgreSQL | Azure Database for PostgreSQL flexible server | | Object storage and notifications | S3 with SQS | Cloud Storage with Pub/Sub | Blob Storage with Event Grid | | Volumes for ClickHouse | EBS gp3 | Persistent Disk (pd-ssd or Hyperdisk) | Azure Managed Disks (Premium SSD) | | Secrets and keys | Secrets Manager and KMS | Secret Manager and Cloud KMS | Key Vault | | Load balancing and TLS | AWS Load Balancer Controller with ACM | GKE Ingress or Gateway with Google-managed certificates | Application Gateway for Containers or NGINX ingress | | LLM endpoint, for example | Amazon Bedrock | Vertex AI | Azure OpenAI | ### LLM providers | Used for | Notes | | - | - | | Policy checks | Classifies a prompt or tool call while the agent waits, so the endpoint's latency matters | | Background reviews | Reviews recorded sessions for findings; latency matters less than throughput | ### Container registry Images are in Asymptote's private registry. Pull them with the credentials Asymptote gives you, or mirror them into your own registry (Amazon ECR, Artifact Registry or Azure Container Registry) and point the chart at it. ## Migration * **From Asymptote Cloud.** Stand up the self-hosted deployment, re-enroll endpoints against it (see [Connecting Endpoints](https://docs.asymptotelabs.ai/self-hosting/connect-endpoints.md)), and disconnect them from Asymptote Cloud. Ask Asymptote about carrying over policies and history. * **From BYOC storage.** Point the self-hosted loader at your existing bucket, and it loads the history already there. ## Next steps * [System Architecture](https://docs.asymptotelabs.ai/self-hosting/system-architecture.md): what runs where. * [Terraform](https://docs.asymptotelabs.ai/self-hosting/terraform.md): create the cluster and services. * [Kubernetes Deployment](https://docs.asymptotelabs.ai/self-hosting/kubernetes.md): install the chart. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.