Back to corpus
architecturetechnical paper candidatescore 66

Voice Ordering Pipeline Architecture

> **Purpose**: Comprehensive technical documentation for the voice ordering system refactoring. > **Last Updated**: December 26, 2025 > **Status**: ✅ Complete (10/10 Steps Done)

Open full HTMLDownload HTMLCorpus JSONPaper schema

Full HTML reader

Read the full artifact

Open in new tab

Extracted abstract or opening context

> **Purpose**: Comprehensive technical documentation for the voice ordering system refactoring. > **Last Updated**: December 26, 2025 > **Status**: ✅ Complete (10/10 Steps Done) 1. [Executive Summary](#executive-summary) 2. [Problem Statement](#problem-statement) 3. [Architecture Overview](#architecture-overview) 4. [Component Details](#component-details) 5. [Data Flow Pipeline](#data-flow-pipeline) 6. [App Routes & Views](#app-routes--views) 7. [File Structure](#file-structure) 8. [Implementation Checklist](#implementation-checklist) 9. [Integration Points](#integration-points) 10. [Technical Decisions](#technical-decisions) 11. [Testing Strategy](#testing-strategy) 12. [Rollback Plan](#rollback-plan) A **modular voice ordering pipeline** that decomposes the monolithic `VoiceOrderingService.swift` (2,464 lines) into 10 focused components: | # | Component | Lines | Responsibility | Status | |---|-----------|-------|----------------|--------| | 1 | FeedbackCoordinator | 337 | TTS + haptics + audio | ✅ DONE | | 2 | SessionManager | 270 | Session lifecycle | ✅ DONE | | 3 | UtteranceCompletionDetector | 407 | Silence/completion detection | ✅ DONE | | 4 | LiveOrderPreviewGenerator | 950 | Real-time item preview | ✅ DONE | | 5 | TranscriptPipeline | 552 | Transcript state management | ✅ DONE | | 6 | CartCoordinator | 440 | Cart state separation | ✅ DONE | | 7 | ConfirmationCoordinator | 626 | Confirmation flow | ✅ DONE | | 8 | OrderParsingPipeline | 1,016 | Hybrid AI+NLU parsing | ✅ DONE | | 9 | VoiceOrderingOrchestrator | 714 | Thin coordinator | ✅ DONE | | 10 | Testing & Cleanup | 600+ | Unit tests (207 tests, 181 passing) | ✅ DONE | 1. **Parsing Strategy**: **Hybrid Merge** - Run AI and NLU parsers in parallel, merge results 2. **Live Preview**: **Pattern Matching** - Fast regex-based matching (~50ms latency) 3. **Implementation Approach**: **Incremental** - Extract one component at a time

Promotion decision

What has to happen next

Promote into a technical note or architecture paper with implementation anchors.

Why this is not always a full paper yet

Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.