feat(ai-autopilot): live bootstrap capstone — real model + LocalRunner (#124)#136
Merged
Merged
Conversation
#124) The live half of the bootstrap capstone. src/live.ts runs the same flow as the offline example but swaps the fakes for real infra: a real model via @gemstack/ai-sdk and a real LocalRunner workspace on the host filesystem. The architect, the build workers, and the deploy decision are all model-driven, and the workers write real files to disk. Verified end to end against Anthropic Haiku 4.5: the architect returned 6 real decisions, three persona workers scaffolded a 9-file Vike + universal-orm orders app (schema + migration, pages/orders/+Page|+data|+config|+Head, a UI-intent renderer), the production-grade loop blocked once on missing auth then cleared, and deploy decided SSR -> dockploy. Scoped for a first bounded proof: the loop keeps the scripted verdict (cheap, deterministic) and deploy uses planOnlyTarget (decides + narrates, does not ship). Booting/serving the generated app + real deploy adapters remain (#109). Adds start:live to the example; offline test + typecheck green.
suleimansh
added
enhancement
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The live half of the bootstrap capstone (#124). The offline example verifies the flow structurally with fakes;
src/live.tsruns the same flow for real.What changed
src/live.ts— swaps the fakes for real infra: a real model via@gemstack/ai-sdk(registers the Anthropic provider fromANTHROPIC_API_KEY) and a realLocalRunnerworkspace on the host filesystem. Architect, build workers, and deploy decision are all model-driven; the workers write real files to disk.src/main-live.ts+start:livescript — runnable entry.Verified end to end
Ran against Anthropic Haiku 4.5 (
pnpm start:live):database/schema.ts,database/migrations/001_create_orders.sql,pages/orders/+Page.jsx/+data.ts/+config.ts/+Head.jsx,lib/ui-intent-renderer.jsx,UI_INTENT_PATTERN.md;LocalRunnerruns on the host, so the only real dependency was a model key — this was never as gated as the issue assumed.Scope (honest boundary)
First bounded, cheap proof: the production-grade loop keeps the scripted verdict (deterministic, no model budget on the checklist), and
deployusesplanOnlyTarget(decides + narrates, does not actually ship). Booting/serving the generated app and real deploy adapters remain, tracked by #109.Private example package, so no changeset. Offline test + typecheck green.