# CareSync AI Discovery Brief Last updated: 2026-08-31 Canonical site: https://care.star-t.io/app ## Canonical summary CareSync is a Korean public demo and pilot-discussion surface for a multilingual senior-care record workflow. It focuses on the gap between foreign care workers' native-language notes, institutional reviewer accountability, and family-facing summaries. CareSync's public demo uses synthetic, de-identified scenarios. It does not process real senior personal data, health data, official care records, voice files, or external AI provider logs. ## Product concept CareSync helps a reviewer inspect a multilingual care note workflow: 1. A care worker or reviewer enters a short visit memo. 2. The system creates a structured draft record. 3. The reviewer compares original memo, transcript, structured fields, and source evidence. 4. The reviewer chooses approve, hold, or request more information. 5. Family preview remains blocked until reviewer approval and consent gates are satisfied. 6. QA guardrails reinforce that the output is not diagnosis, prescription, emergency judgment, or an automatically finalized official record. ## Current public demo Demo URL: https://care.star-t.io/app/demo The demo includes six synthetic multilingual care-note cases in an institutional review queue. A user can select cases and change queue states through approve, hold, and request-more-information actions. KPI cards show pending cases, approved cases, blocked cases, and estimated time savings for the synthetic workflow. ## Education Record Lab Record Lab URL: https://care.star-t.io/app/training Record Lab is the education-first synthetic practice surface for foreign caregiver candidates and instructors. It uses four built-in multilingual scenarios. Learners separate observations explicitly present in a memo from fields that require human confirmation; the deterministic rubric then shows matched, missed, and unnecessary selections. The local summary is not stored and is not a qualification, clinical, language, cohort, or production-AI score. ## Intended audiences - Senior-care education institutions - Foreign resident and multicultural support institutions - Long-term-care institutions with reviewer workflows - Public-sector and local-government pilot partners - Families and operators evaluating safe family-preview boundaries ## Safety and claim boundary CareSync's public site should be described as: - A demo and pilot-discussion surface - A synthetic-data workflow demonstration - A reviewer-in-the-loop care-record draft workflow CareSync's public site should not be described as: - A medical device - A diagnostic AI - A prescription or emergency decision tool - A production system that already handles real senior data - An automatic official care-record generator - An automatic family-message delivery system ## Public pages - Home: https://care.star-t.io/app - Operational demo: https://care.star-t.io/app/demo - Education Record Lab: https://care.star-t.io/app/training - Pilot inquiry: https://care.star-t.io/app/contact - Privacy and claim boundary: https://care.star-t.io/app/privacy - Pilot method: https://care.star-t.io/app/pilot-method - Safety guide: https://care.star-t.io/app/safety-guide - Workflow comparison: https://care.star-t.io/app/record-workflow-comparison - Synthetic evidence case: https://care.star-t.io/app/synthetic-evidence-case - Insights hub: https://care.star-t.io/app/insights - Roundtable entry: https://care.star-t.io/app/community - Evidence and reports: https://care.star-t.io/app/reports - Content feed: https://care.star-t.io/feed.xml ## Scheduled owned-site insights - 기관 검토자의 수정은 왜 근거와 함께 남아야 하는가: https://care.star-t.io/app/insights/reviewer-edit-evidence Scope: Synthetic demo workflow note about reviewer evidence, edits, and hold/approval states; it does not claim a deployed institution workflow or ERP integration. ## Citation-ready paragraph CareSync is a multilingual senior-care record review workflow for pilot discussions. Its public demo uses synthetic multilingual care-note cases to show how native-language notes can become Korean structured draft records, how institutional reviewers can approve, hold, or request more information, and how family previews remain blocked before approval and consent. ## Evidence available on the public site - The home page explains the problem, workflow, pilot boundary, and FAQ. - The demo page shows a synthetic institutional review queue and reviewer decisions. - The Record Lab page supports source-grounded synthetic record-writing practice for learners and instructor discussion without accounts, free-form case collection, or real care data. - The privacy page states that the public demo does not collect or process real senior personal data, health data, or sensitive care data. - The pilot-method page separates public-source context, current demo evidence, and blocked production capabilities. - The safety-guide page explains the no-real-data, human-review, and no-auto-share boundaries. - The workflow-comparison page describes CareSync as an ERP-before, reviewer-facing input layer rather than an ERP replacement or competitor ranking. - The synthetic-evidence-case page reports deterministic fixture regression evidence and explicitly excludes field, clinical, customer, and real-data performance claims. - The contact page uses mail draft handoff rather than server-side lead storage. - The insights hub curates the public pilot, safety, workflow, and synthetic-evidence materials. The roundtable entry offers role-based reading paths for institutions, education providers, and prospective care workers; neither page collects member posts, care cases, or community account data. - The reports hub separates official public sources, synthetic repository evidence, consent-gated institution research, and future account-connected analytics. It currently publishes only synthetic-evaluation and pilot-readiness reports, with no automatic external collection, real care data, or field-performance claims. ## Remaining external verification The public files support crawler and AI discovery. Search Console submission, Bing/Naver webmaster receipts, and actual search or AI citation evidence require external account verification and are not proven by the repository alone.