One engineer can install everything on the front page and use all of it. A team pays when the definitions, the record and the sign-in have to be shared, or when they want the DistillQL LLM to do the asking.
| Open$0, Apache 2.0One engineer, one laptop or server. | Team$900 a monthA data team that owns the company's definitions. | Inference$900 a month, plus tokensTeam, plus the DistillQL LLM: private and low cost by design. | EnterpriseTalk to usSeveral teams, several warehouses, a security review. | |
|---|---|---|---|---|
| The engine | ||||
| Definitions | One file, drafted or verified | Shared, versioned, signed off | Shared, versioned, signed off | Approval rules per domain |
| The agent | Yours, over MCP or Python | Yours | The DistillQL LLM. Private and low cost by design; no Claude or Cursor needed | Yours or ours |
| The record | Returned with each answer | Stored and searchable | Stored and searchable | Retention you set, SIEM export |
| Sync | dbt, Semantic Views, catalogs | dbt, Semantic Views, catalogs | Plus a catalog you name | |
| Sign-in | Own connection or role | SSO mapped to roles | SSO mapped to roles | SSO, SCIM, your own key |
| Databases | Unlimited | Up to five | Up to five | Unlimited |
| People asking | Unlimited | Unlimited | Unlimited | Unlimited |
| Support | GitHub issues | Email, one business day | Email, one business day | A named engineer |
While DistillQL is pre-release, Team is free for design partners, who keep that price for a year after 1.0. Inference is billed by the token; the DistillQL LLM runs on private inference that keeps no logs, and we pass the cost through with a fifth on top. The open package makes no model calls of its own, on any tier.