Skip to content
Almathal Handbook

Version Roadmap

Accepted

This document sketches the version progression of Almathal beyond MVP. It is intentionally approximate: the right pace is set by what real customers need, not by a pre-built schedule.

For exact MVP scope, see MVP Scope.

v0 — Pre-MVP Design Phase

Current. The architecture, schemas, and decision log are produced. No code yet.

v0.x — MVP

Goal: Prove the architecture and ship to first enterprise pilot.

Includes:

  • Two Archetypes (RAG Chatbot, CRUD Admin Tool)
  • ~150 Component Library Adapters
  • ~5–8 Modules
  • Full pipeline implementation
  • Trust infrastructure (provenance, license, CVE, audit trail)
  • Stage 1–2 Curation Pipeline
  • Customer web UI

v1 — Production Maturity

Goal: Address the gaps that prevent broader enterprise adoption.

Archetypes added

  • Microservices architectures (per ADR-0015)
  • Document processing pipeline
  • AI agent / workflow automation
  • Internal analytics dashboard (Java + Python + React, cross-language)
  • API-only services (no frontend)

Component Library expanded

  • Cloud-vendor SDKs (AWS, GCP, Azure)
  • Additional Java enterprise (Spring Batch, Camunda, Mule)
  • Mobile frontend exploration (React Native first, if demand exists)
  • Expanded AI/LLM coverage (more providers, more orchestration)

Target: ~300–400 Adapters total.

Module Library expanded

  • Payments Module (Stripe, Adyen)
  • Billing Module (Lago wrapper or platform-composed)
  • Email/notification Module
  • File upload and storage Module
  • Search Module (Elasticsearch/OpenSearch-based full-text)
  • Workflow / approval engine Module

Target: ~20–30 Modules.

Pipeline features

  • Exploratory Archetype path (replaces pure-LLM fallback for unmatched requests)
  • Cross-build re-generation with customization preservation
  • Multi-app orchestration for microservices Archetypes
  • CI/CD integration for auto-regeneration

Trust infrastructure

  • Optional verified artifact mirror (per ADR-0017) for enterprise tier
  • On-prem and air-gapped deployment options
  • SOC 2 Type II audit completion
  • Customer-specific Manifest extensions (scoped, not global)
  • Stage 3 Curation Pipeline enabled for low-risk categories

Customer surface

  • API access (machine-driven generations)
  • Multi-user team collaboration on a Spec
  • IDE plugins (start with VS Code)

Compliance packs

  • HIPAA-ready Archetype variants
  • PCI-DSS-ready Archetype variants

v2 — Self-Improving Platform

Goal: The Curation Pipeline operates at scale with the LLM in expanded curator and authorship roles. The library grows faster than the team alone could grow it.

Curation evolution

  • Agent-driven Adapter discovery scales beyond what human-only could match
  • LLM authorship of Manifests with structured human ratification across all non-security categories
  • Pattern discovery: agents observe successful and failed generations across customers, propose new patterns and Archetype candidates
  • Closed-loop learning from production: real outcomes feed Adapter/Module quality scores

Cross-customer signals (with privacy preservation)

  • Aggregate usage patterns inform Component Library prioritization
  • Anonymized failure patterns inform Stitcher prompt improvements
  • Customer-opt-in feedback channels for Adapter quality

Marketplace

  • Community Adapter contributions (with strict review)
  • Customer-published Modules (scoped to their organization or opt-in public)
  • Third-party Archetype contributions

Language slice expansion

  • Go slice
  • C# / .NET slice (a major Java alternative for enterprise)
  • Rust slice (where it has enterprise traction)

v3+ — Open Questions

These are real possibilities but with too many unknowns to commit:

  • Could the system handle existing-codebase modernization (ingest a legacy system, model it as Modules + customizations, then improve)?
  • Could the platform serve general-population developers via a Cursor-style IDE integration, leveraging the same Component Library?
  • Could it generate apps for embedded or edge-compute scenarios?
  • Could it become an upstream contributor of canonical Manifests to scaffolder communities (JHipster, etc.)?

Decisions on these are deferred until v1 reality is settled.

Cadence Targets

The cadence we aim to sustain:

  • Patch releases: weekly, primarily security-driven Manifest updates
  • Minor releases: monthly, including new Adapters, Modules, and Archetype refinements
  • Major releases: quarterly to half-yearly, including schema evolution and breaking improvements

Indicators That Will Re-shape This Roadmap

Real-world signals will adjust priorities:

  • Which Archetypes customers ask for that don’t exist yet
  • Which Adapters break or get deprecated upstream
  • Which platform features customers refuse to buy without (procurement-driven)
  • Where the LLM-stitching token economics actually land relative to projection
  • Whether the Curation Pipeline keeps up with upstream ecosystem velocity