Costs
Lakehousecat deployments involve three cost categories. In most production environments, API and LLM provider costs are the dominant cost driver, followed by infrastructure, with subscription fees being the smallest component. All three scale with usage and user growth.
For subscription tiers, pricing, and seat/instance limits, see Billing.
1. API and LLM Provider Costs
Lakehousecat integrates with external AI providers to power chat, analytics, and speech-to-text functionality. These costs are typically the largest variable in a Lakehousecat deployment and scale directly with usage.
Supported providers:
| Provider | Capabilities |
|---|---|
| OpenAI | Chat, Speech-to-Text |
| Anthropic | Chat |
| Chat, Speech-to-Text | |
| Azure (Azure OpenAI) | Chat, Speech-to-Text |
| AWS (Bedrock / Transcribe) | Chat, Speech-to-Text |
| xAI | Chat, Speech-to-Text |
Customer Responsibility
The API keys used to call these providers belong to the customer. This means:
- The customer registers their own API credentials with each provider
- Costs for API calls (tokens processed, audio transcribed) are billed by the provider directly to the customer
- Customer data — including parts of queries and data source content — is sent to the configured providers as part of normal operation
- Lakehousecat does not monitor or limit API spending on the customer's behalf
Provider pricing structures change frequently. Customers are responsible for monitoring their own API usage and costs using the provider's native tooling (e.g., OpenAI usage dashboard, AWS Cost Explorer).
As user count and query volume grow, API costs grow proportionally. High-concurrency scenarios (many users querying simultaneously) can produce significant API spend. Customers should configure provider rate limits and spending alerts to avoid unexpected charges.
2. Infrastructure Costs
Lakehousecat is deployed as a Kubernetes Operator on the customer's own Kubernetes cluster. Lakehousecat provides the Operator software — the underlying infrastructure is fully customer-managed.
This means:
- Node type and size are chosen by the customer's cluster administrator, not by Lakehousecat
- Compute, storage, and network costs are billed directly by the customer's cloud provider or on-premises infrastructure
- Scaling decisions (horizontal and vertical) are made by the customer via the Operator configuration
- Lakehousecat provides the Operator and ensures compatibility with standard Kubernetes environments
Infrastructure costs scale with the number of users and the data volume processed. Infrastructure cost optimization (right-sizing, auto-scaling, reserved instances) is the responsibility of the cluster administrator.
3. Subscription and User Costs
Subscription tiers, pricing, and seat-based billing mechanics have moved to their own section — see Billing for the tier table and pricing example, and Seats & Instance Limits for seat billing rules (including what happens to deactivated or Pending users).