From structured intent to governed release

Keewo compiles business requirements into Vue 3 + Express TypeScript applications through a 6-step governed pipeline. In practice, that delivery model combines a customer-facing runtime, a controlled Lab experience for builders, and a governed release path.

Feature requirements

Capture business intent as a bounded feature slice.

AI draft plan and contract

Use AI to draft a technical plan without skipping governance. The AI-assisted plan must preserve required model coverage, identifier quality, and localization intent before it can become governed source material.

Managed registry

Review and store the governed model for entities, views, workflows, rules, security, integrations, and patterns.

Generated XML modules

Generate inspectable modules for data, logic, UI, and security. Each feature produces 6 files (Data, Security, Language, Logic, Tests, UI) plus hash-verifiable provenance that connects output to the governed source model.

Compile, validate, test, and build

The framework compiles XML modules into Vue 3 frontend and Express backend TypeScript applications. Quality evidence covers validation, tests, business rules, workflow checks, migration safety, authoring gates, and release readiness.

Human approval and promotion

Prepare the change for review, approve it through an accountable human decision, promote the baseline, and hand the result off to source control.

Runtime, Lab, and Managed Authoring

Keewo is not only a generated application runtime. It is an operating model with three distinct surfaces:

Runtime: the deployed application for business users
Lab: the builder surface for inspectable editing, generation, testing, and preview
Managed Authoring: the governed path for registry review, materialization, validation, promotion, and release

Why this matters

AI contribution stays bounded by Managed Authoring and guardrail validation.
Generated artifacts remain inspectable, versioned, and hash-verified.
Release governance is explicit: quality evidence, approval, promotion, and provenance travel with the change.

How pilots are evaluated

Keewo pilots are designed to let a customer evaluate both the runtime and the platform story. Business users validate the runtime. Technical evaluators review Keewo Lab in a controlled sandbox. Production promotion stays governed.

What the pipeline produces

Input: managed registries define entities, screens, workflows, business rules, security rules, integrations, and application patterns.

Intermediate: inspectable modules and hash-verified provenance tie generated output back to the governed model.

Output: Vue 3 frontend (TypeScript, vue-router, Vite) + Express backend (TypeScript, Prisma ORM). Standard frameworks, no proprietary runtime.

Concrete Artifacts

Sanitized artifacts from a sample service desk change

A bounded change request updates the managed model, materializes output, and carries quality evidence into approval.

Change Request

request:
  kind: change_request
  goal: highlight high-priority service work
  scope:
    - dashboard metric
    - service request workflow board
    - business rule evidence
  reviewer: accountable human approver

Managed Model Excerpt

model:
  entities:
    - service request
    - policy account
    - service owner
  workflow:
    states:
      - intake
      - review
      - waiting for customer
      - completed
  dashboard:
    cards:
      - high-priority requests
      - waiting for customer

Quality Evidence

readiness:
  validation: passed
  tests: passed
  businessRules: passed
  workflowChecks: passed
  migrationSafety: passed
  authoringGates: passed
governance:
  approval: required
  promotion: human decision

Output: Vue 3 frontend + Express backend, both TypeScript. The public site shows sanitized excerpts only; complete artifacts remain reviewable inside the governed project workspace.