Security & Architecture
Alethiom deploys in your cloud account. We operate it through a role you grant.
This page enumerates what runs in your account, what moves and what doesn’t, and the controls that stay in your hands: the logs you can read, the access you can revoke.
Data flow
No data from your warehouse or the serving layer is stored outside your cloud account.
The whole product runs inside your account, so query results go from your cloud straight to your users’ browsers.
data plane
Event sourcesvendor exports · CDP · server-side streams — however your pipeline works today
your cloud account
Your warehousesystem of record
Alethiom deploymentengine · UI · APIs
sync workers
Your team’s browsersquery results, passed securely
what moves: bounded, scheduled sync queries from warehouse to serving layer, using your warehouse’s native CDC capabilities (inside the account)
what never moves: customer event data, to Alethiom
Deployment model
One stack, one role, revocable at any time.
- You run one infrastructure stack from us, and it creates one scoped role in your account.
- Through that role, we build and operate the Alethiom stack inside your account: provisioning, upgrades, monitoring, and incident response. You never have to operate our software; you just host it.
- You can revoke the role at any time. That cuts our access immediately.
- Ships as
- A standard Kubernetes cluster running the full stack, visible to your own audit and security tooling.
- Warehouse access
- Snowflake, BigQuery, and Databricks supported today. A read-only role you grant, scoped to the datasets you choose. Only scheduled CDC sync queries touch the warehouse as often as every 5 minutes, supporting record-level appends, edits, and deletes. Interactive queries never hit your warehouse.
- Operations
- Your cloud’s audit logs record any infrastructure action we take through the role.
- Encryption
- TLS in transit; at rest under your cloud’s storage encryption.
The stack
What runs in your account.
One Kubernetes cluster in your account runs four things:
- ClickHouse
- The query engine: open-source ClickHouse, holding a disposable cache built from your warehouse and rebuildable at any time.
- Postgres
- Metadata: definitions, boards, configuration, and AI conversation history, including the report results a conversation showed.
- Keycloak
- Sign-in, federated to your identity provider: your SSO, your MFA policy, your deprovisioning. No separate password store.
- Application services
- The UI, the APIs, the AI service, and the sync workers.
Connections
Required network connections.
- Alethiom telemetry
- Usage telemetry, so we can make sure the product is running well, plus feedback your users choose to send.
- Anthropic API (with AI enabled)
- AI calls to Anthropic. Data is never used for training purposes. One switch turns AI off entirely, or bring your own Anthropic key.
- MCP Connector
- Users sign into an MCP client as themselves and run read-only reports under their own permissions.
- Alerts
- Outbound alerts to Slack or Microsoft Teams.
AI posture
One switch turns it off.
AI is part of every deployment. If your policy needs it off, one switch turns it off.
AI off
Zero model-provider traffic, and MCP connections are disabled. Everything else in the product works exactly the same.
AI on
Model calls run under Anthropic’s commercial API terms: your data is never used for training purposes.
You can bring your own Anthropic key, and AI runs under your Anthropic API credentials.