{"id":51475373,"url":"https://github.com/asmuelle/sitescribe","last_synced_at":"2026-07-06T20:30:33.996Z","repository":{"id":365993003,"uuid":"1266479435","full_name":"asmuelle/sitescribe","owner":"asmuelle","description":null,"archived":false,"fork":false,"pushed_at":"2026-06-19T19:52:33.000Z","size":97,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-06-19T20:20:47.039Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Swift","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/asmuelle.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2026-06-11T16:50:39.000Z","updated_at":"2026-06-19T19:52:37.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/asmuelle/sitescribe","commit_stats":null,"previous_names":["asmuelle/sitescribe"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/asmuelle/sitescribe","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmuelle%2Fsitescribe","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmuelle%2Fsitescribe/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmuelle%2Fsitescribe/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmuelle%2Fsitescribe/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/asmuelle","download_url":"https://codeload.github.com/asmuelle/sitescribe/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmuelle%2Fsitescribe/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35205739,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-07-06T02:00:07.184Z","response_time":106,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-07-06T20:30:33.292Z","updated_at":"2026-07-06T20:30:33.985Z","avatar_url":"https://github.com/asmuelle.png","language":"Swift","funding_links":[],"categories":[],"sub_categories":[],"readme":"# SiteScribe\n\n[![CI](https://github.com/asmuelle/sitescribe/actions/workflows/ci.yml/badge.svg)](https://github.com/asmuelle/sitescribe/actions/workflows/ci.yml)\n\n\u003e Offline AI field-documentation copilot: photograph, dictate, and auto-generate structured inspection reports, receipts, and job records at zero signal, syncing when coverage returns.\n\n**Category:** Edge AI / on-device inference (iOS + Android) ·\n## Concept\n\nOffline AI field-documentation copilot: photograph, dictate, and auto-generate structured inspection reports, receipts, and job records at zero signal, syncing when coverage returns.\n\n## Target User\n\nHome/building inspectors, insurance adjusters, utility and telecom field techs, and solo trade contractors — workers whose sites (basements, rural routes, industrial buildings) are dark zones for entire shifts and who bill by the completed report. Solo-contractor receipt/invoice capture is the bottom-up wedge into firms.\n\n## Why Edge AI Is Structural (not decoration)\n\nAFM 3 image input + OCRTool + BarcodeReaderTool read equipment tags, serial plates, meter values, and receipts from photos; @Generable fills the user's report schema (defect, location, severity, recommended action) deterministically; whisper.cpp transcribes walk-through voice notes offline; Spotlight local RAG answers 'what did I flag at this property last visit?' across past reports and scanned documents. Android parity via Gemma 4 E2B structured output in the AICore Developer Preview. Essential: 30% of frontline workers explicitly expect offline-capable apps — every cloud copilot returns a spinner exactly when the work happens, and unlimited daily extraction is free on-device where cloud-vision rivals bleed margin.\n\n## Why Now (2026 timing)\n\nMobile document scanning + local LLM structured extraction has 'no polished commercial player' (only OSS hobby apps), and field ops/OCR is cited as the strongest near-term on-device use case; @Generable, OCRTool, and local RAG all shipped as free system primitives at WWDC26 — assembly cost just collapsed for whoever moves first.\n\n\n## Tech Stack\n\niOS (Swift/SwiftUI, iOS 26 floor with iOS 27 fast-follow): VisionKit DataScannerViewController + Vision RecognizeDocumentsRequest as the universal OCR/barcode path (works on all recent iPhones); FoundationModels @Generable guided generation against the user's report schema, using AFM 3 Core (text) everywhere and AFM 3 Core Advanced + OCRTool/BarcodeReaderTool only on capable devices; SpeechAnalyzer/SpeechTranscriber for offline dictation with whisper.cpp (large-v3-turbo, Metal) as fallback on older devices; NLContextualEmbedding + GRDB/SQLite with sqlite-vec for local RAG over past reports; BGTaskScheduler sync queue with an optional cloud polish pass (Claude Haiku-class or Gemini Flash via Foundation Models' new third-party adapter) on reconnect; PDFKit for report output. Android (Kotlin/Jetpack Compose): ML Kit Document Scanner + Text Recognition v2 for OCR (runs on ALL devices, no AICore dependency — this is the load-bearing choice); ML Kit GenAI Prompt API with structured output on Gemini Nano-capable flagships; MediaPipe LLM Inference API (LiteRT) with bundled Gemma 3n E2B + constrained decoding as the mid-tier fallback, accepting the ~2-3GB model footprint, and cloud-deferred extraction on low-end devices; whisper.cpp (or ONNX Runtime Mobile Whisper) for ASR; Room + sqlite-vec for local RAG; WorkManager sync queue. Shared: schema-first report templates (JSON Schema source of truth compiled to @Generable structs and Kotlin data classes), offline-first store with CRDT-ish merge on sync, and CSV/PDF/Spectora-importable export formats.\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fasmuelle%2Fsitescribe","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fasmuelle%2Fsitescribe","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fasmuelle%2Fsitescribe/lists"}