Skip to main content

Roadmap V2 Reorganization

ARCHIVE — This document is historical reference only. It may contain outdated information. See docs/status.md for current project state.

Date: February 16, 2026 Status: APPROVED Impact: Major architectural reorganization

Context​

The original roadmap evolved organically from Phase 1 → Phase 3p with 16 sub-phases (3a-3p). As the project matured and research deepened (OpenMorph, WASM extensions), it became clear that the phase numbering obscured the architectural relationships between features.

User motivation: "I have 3 kids" — multi-modal identity (voice/face auth for kids without email) is a core use case, not a late-stage feature.

Decision​

Reorganize from numbered phases to version-driven layers that reflect architectural dependencies and product vision.

Changes​

Phase Consolidation​

Old PhaseNew LayerVersionRationale
3e.1-3e.5 (UX)Layer 2V0.9Product completeness
3j (Settings as Integration)Layer 2V0.91-week effort, merge with completeness
3i (OpenMorph)Layer 3V1.0Revolutionary, git-native, defines platform
3l (WASM Extensions)Layer 4V1.2Extensibility, rebuild integrations as WASM
3f (Professional Integrations)Layer 4V1.2JIRA/Notion/GitHub as WASM extensions (validates system)
3o (Multi-modal Identity)Layer 5V1.0-1.3Kids without email, voice/face auth - CORE PERSONA
3p (Video) + 3k (Interactive UI)Layer 5V1.3Merged: Multimodal Interaction
3d M3 (Full Offline Mobile)Layer 6V1.5Depends on OpenMorph + WASM
3h (Rich Content Views)Layer 7V2.0Production polish
3n (Self-Aware AI)Layer 8V2.0+Advanced/optional
3m (Crypto Marketplace)Layer 8V2.5+Deprioritized - optional module

Strategic Shifts​

  1. WASM-First Strategy

    • Freeze new hardcoded integrations (no more Python backend/interfaces/integrations/*.py)
    • Build JIRA/Notion/GitHub as WASM extensions
    • Validates extension system with real use cases
    • Same binary runs on Python backend + Tauri frontend (zero duplication)

  2. OpenMorph as V1.0 Defining Feature

    • .morph/ directories (like .git/) make directories AI-augmentable
    • Everything lives in Git: tasks, skills, conversations, configs
    • Database becomes read cache
    • Portable Spaces (.morph/ = Space)
    • Breaking change acceptable (not in production yet)

  3. Multi-Modal Identity for Families

    • Kids without email can authenticate via voice/face ("Morphee, it's me")
    • Accessibility (elderly, visually impaired)
    • Core persona (Sophie, age 8) requires this
    • Move from Phase 3o (post-V2.0) to Layer 5 (V1.0-1.3)

  4. Crypto Marketplace Deprioritized

    • Interesting vision but distracts from core product
    • Can be built as WASM extension later
    • Move to Layer 8 (Future/Optional)

Version Milestones​

V0.9 (Layer 2)   — Product Readiness (2 weeks)
└─ Complete audits (35% remaining)
└─ Settings as Integration
└─ Branding/logo
└─ FAB quick actions

V1.0 (Layer 3)   — OpenMorph Revolution (8 weeks) 🚀
└─ Git-native architecture
└─ .morph/ directories
└─ Tasks/Skills/Configs in Git
└─ Branching, temporal navigation
└─ Multi-sync modes
└─ BREAKING CHANGE: migration required

V1.2 (Layer 4)   — Extension Ecosystem (12 weeks)
└─ WASM runtime (wasmtime-py, wasmtime-rs, wasm3)
└─ OCI registry (rg.morphee.ai)
└─ JIRA/Notion/GitHub as WASM extensions
└─ Marketplace UI
└─ Code signing, permissions

V1.3 (Layer 5)   — Multimodal Interaction (8 weeks)
└─ Voice/face authentication
└─ Kids without email
└─ Video/gesture recognition
└─ Interactive/editable components
└─ Haptic feedback

V1.5 (Layer 6)   — Full Offline (6 weeks)
└─ Local LLM (candle, small models)
└─ Mobile offline (3d M3)
└─ Desktop offline
└─ Sync engine
└─ Conflict resolution

V2.0 (Layer 7)   — Production Platform
└─ Monitoring (PostHog, Grafana)
└─ Channels (WhatsApp, Telegram)
└─ Browser extension
└─ Partner framework

V2.5+ (Layer 8)  — Future/Optional
└─ Self-aware AI (codebase as memory)
└─ Crypto marketplace (if demand exists)

Unified Capability Model​

Old model had 4 overlapping concepts:

  • BaseInterface (Python classes)
  • Skills (database-stored workflows)
  • WASM Extensions (portable code)
  • Frontend Components (dynamic UI)

New unified model - everything is a "Capability":

# .morph/capabilities/send_email.yaml
type: wasm_extension
binary: .morph/extensions/gmail.wasm
actions:
  - name: send_message
    permissions: [http:write, data:read]

# .morph/capabilities/daily_briefing.yaml
type: skill
steps:
  - action: tasks__list_pending
  - action: google_calendar__list_events_today
  - action: frontend__show_card

Capability types:

  1. WASM Extension - portable code (.wasm binaries)
  2. Skill - workflow (YAML step sequences)
  3. UI Renderer - dynamic components (ComponentSpec)
  4. LLM - reasoning engine (special capability)
  5. Memory - knowledge storage (special capability)

Implementation Timeline​

Immediate (Week 1-2): V0.9 RC​

  • Complete audit fixes (35% remaining)
  • Implement Settings as Integration (3j - 1 week)
  • Branding/logo (M-DESIGN-003)
  • FAB quick actions (L-IDEA-001)
  • Tag V0.9 Release Candidate

Next (Week 3-10): V1.0 OpenMorph​

  • Follow 2026-02-16-git-native-spaces-openmorph.md
  • 3i.1: Spec + Branching (2 weeks)
  • 3i.2: Migration (3 weeks) - BREAKING CHANGE
  • 3i.3: Sync Layer (2 weeks)
  • 3i.4: Polish (1 week)
  • Tag V1.0 - Git-Native, Portable

Then (Week 11-22): V1.2-1.3 Parallel​

Team A: WASM Extensions (12 weeks)

Team B: Multimodal (8 weeks)

  • Voice/face enrollment
  • Kids without email flow
  • Video/gesture recognition
  • Step-up authentication

Benefits​

  1. Clearer vision - version numbers signal product evolution
  2. Architectural coherence - OpenMorph influences everything
  3. Community-ready - WASM extensions enable third-party integrations
  4. Family-friendly - kids without email is V1.0, not optional
  5. Portable by design - .morph/ directories are the unit of portability
  6. Git-native - version control for AI agents
  7. Offline-first - server is sync hub, not dependency

Breaking Changes​

V1.0 Migration (Acceptable - Not in Production)​

What breaks:

  • Tasks table → .morph/tasks/*.md (PostgreSQL → Git)
  • Skills table → .morph/skills/*.yaml (PostgreSQL → Git)
  • Interface configs table → .morph/config.yaml (partially PostgreSQL → Git)

Migration script:

python scripts/migrate_to_openmorph.py --group-id=<id>
# Exports PostgreSQL → .morph/ directory per Space
# Database becomes read cache

User impact:

  • One-time migration per Space
  • Existing data preserved (exported to Git)
  • New superpowers: branching, temporal navigation, portability

Open Questions​

  1. Phase 3e.6 (Local/Remote Task Execution) - Approved design exists. Should this be:

    • Part of V1.0 (OpenMorph) - task execution metadata in .morph/?
    • Separate V1.1 feature?
    • Deferred until V1.5 (Offline layer)?

  2. Slack Integration (3f partial) - MVP exists. Should we:

    • Keep Python implementation for now?
    • Rebuild as first WASM extension (validates system)?

  3. Multi-modal identity timing - V1.0 or V1.3?

    • V1.0: Kids can use app immediately (core persona)
    • V1.3: Parallel with WASM extensions (12 weeks vs 8 weeks)

References​

Approval​

✅ Approved by Sebastien Mathieu - February 16, 2026

All documentation updated to reflect Layer/Version model:

  • docs/ROADMAP.md
  • docs/status.md
  • CLAUDE.md
  • MEMORY.md