Migrated from the archived repo: Database-Tycoon/localhost-stack — a batch of four product-vision notes filed together on 2026-04-04 (#16, #17, #19, #20), consolidated here verbatim so nothing is lost when the old repo is dropped. Much of this vision has since been formalized in the control-plane proposal (#79) and the ingestion rewrite project (#81); retriage for anything not yet captured there.
localhost-stack#16 — Strategic Vision: Transition from Toolset to Unified Data Engine
The current command structure (e.g., 'tycoon data sources add') is fragmented and requires the user to manage the pipeline hierarchy manually. We should move toward action-oriented verbs that drive a state machine (e.g., 'tycoon sync github' which handles adding, ingesting, and transforming in one unified flow). This reduces cognitive load and aligns with the mission of zero-config analytics.
localhost-stack#17 — Architecture: Unified Configuration and Stateful Execution
The CLI should act as a transparent window into the 'tycoon.yml'. Changes to the configuration (e.g., adding a new source in YAML) should be automatically detected by the CLI during execution, triggering only the necessary delta-updates (stateful execution). The goal is to make the terminal and the YAML file feel like a single, unified interface.
localhost-stack#19 — Developer Experience: Deep Integration of AI into the Command Workflow
Move away from 'tycoon ai' as an add-on command and integrate AI capabilities directly into the core commands. The engineer should be able to use natural language intent (e.g., 'tycoon sync my-github-repo') rather than manually chaining discrete pipeline steps.
localhost-stack#20 — Product Vision: Achieving the 'Single Primitive' Goal
To fulfill the mission of local-first, zero-config analytics, tycoon must evolve from a collection of integrated tools (dlt, dbt, Rill) into a single, unified primitive. The user should interact with the 'tycoon' engine, treating the underlying components as implementation details that are managed automatically by the system.
Coverage at migration (2026-08-09)
Mirrored in Jira: PTC-117 (unparented, pending roadmap call). Dev work continues here.
localhost-stack#16 — Strategic Vision: Transition from Toolset to Unified Data Engine
The current command structure (e.g., 'tycoon data sources add') is fragmented and requires the user to manage the pipeline hierarchy manually. We should move toward action-oriented verbs that drive a state machine (e.g., 'tycoon sync github' which handles adding, ingesting, and transforming in one unified flow). This reduces cognitive load and aligns with the mission of zero-config analytics.
localhost-stack#17 — Architecture: Unified Configuration and Stateful Execution
The CLI should act as a transparent window into the 'tycoon.yml'. Changes to the configuration (e.g., adding a new source in YAML) should be automatically detected by the CLI during execution, triggering only the necessary delta-updates (stateful execution). The goal is to make the terminal and the YAML file feel like a single, unified interface.
localhost-stack#19 — Developer Experience: Deep Integration of AI into the Command Workflow
Move away from 'tycoon ai' as an add-on command and integrate AI capabilities directly into the core commands. The engineer should be able to use natural language intent (e.g., 'tycoon sync my-github-repo') rather than manually chaining discrete pipeline steps.
localhost-stack#20 — Product Vision: Achieving the 'Single Primitive' Goal
To fulfill the mission of local-first, zero-config analytics, tycoon must evolve from a collection of integrated tools (dlt, dbt, Rill) into a single, unified primitive. The user should interact with the 'tycoon' engine, treating the underlying components as implementation details that are managed automatically by the system.
Coverage at migration (2026-08-09)
Mirrored in Jira: PTC-117 (unparented, pending roadmap call). Dev work continues here.