Prompt-to-app builders start from a blank database. Mako's Apps start from yours: synced sources, dbt models, and saved queries are already bound when the first component renders.
Live miniature, open on the account-health app — try the cohort filter in the preview, or read how the agent built it in the chat.
Some tools need input, not just charts: an account review app where CS can leave notes, an ops console with approve/reject buttons, an internal tool with filters and drill-downs. Apps are real React — components, state, interactions — versioned and published to the workspace, not a grid of tiles.
Generic app builders make you bring your own backend and hope the schema in the prompt matches reality. Every Mako app has a __datasources folder binding it to the models and saved queries underneath — the same schema the agent queries every day, with auth handled by the workspace.
“Add a cohort filter.” “Make the risk score sortable.” “Only ops can see the refund tab.” The agent edits the app the way it edits SQL — with your schema and conventions in context — next to a live preview. Every change builds, versions, and waits for Publish.
“We replaced a pipeline vendor, an orchestrator, a BI tool, and two desktop licenses with one docker-compose file. The council was impressed.”
Describe it in a sentence. The data layer is already there.