ORVYCOMPANION · v0.1
ORVY Companion›Product Charter
PHASE 0 · WORKING MANUSCRIPT

ORVY Companion — Product Charter v0.1

EXPLORATORY DRAFTNOT CANONICALv0.1 · 18 OCT 2026

Mission, personas, MVP contract and review gates.

Authority statement: This document is a proposal for ORVY Companion; it is not an approved ORVY Core architecture change. Application implementation is not yet authorized.

Document ID: ORVY-COMP-CHTR-001 Status: EXPLORATORY WORKING DRAFT — NOT APPROVED / NOT CANONICAL Prepared: 2026-10-18 Context: Create AI Agents competition, Personal AI Agent category Parent vision: ORVY — Human-first Amplified Intelligence

1. Executive intent

ORVY Companion is an opt-in, context-aware personal AI agent that connects a person's interests, immediate needs, and permissioned location with verified opportunities in the physical world. It seeks to assist with discovery, learning, travel, and practical everyday choices before a user has formulated a search query. It remains advisory by default. The person controls permissions, memory, recommendations, and consequential actions.

Product proposition: ORVY doesn't just answer what you ask. It helps you discover what matters to you — wherever life takes you.

2. Why this product exists

Useful places, activities, and services are frequently overlooked because people may not know they exist, may not realize why they are personally relevant, or may not think to search at the right moment. Conventional search answers a known query. ORVY Companion explores permissioned context and surfaces timely, relevant opportunities with evidence and an understandable next action.

3. Product identity and authority

The working public name is ORVY Companion, development identifier ORVY 0.1. This is a candidate first ORVY Amplifier, not an implementation or replacement of ORVY Core. ORVY Core constitutional, architectural, provenance, and admission rules remain independently authoritative. Findings from this prototype may inform later proposals but never automatically alter canonical ORVY Core decisions.

4. Initial user personas

  • Curious explorer: enjoys culture, heritage, museums, and local learning.
  • Everyday planner: seeks useful nearby services and efficient daily choices.
  • Wellbeing-conscious user: wants general informational guidance and reliable nearby pharmacies or clinics, without unsupervised diagnosis or prescription.
  • Privacy-conscious user: expects visible permissions, explanation, suppression controls, and memory review.

These are initial test personas rather than claims about users or demographic groups.

5. MVP promise and essential capabilities

  1. A responsive, mobile-first application with user sign-in and clear onboarding.
  2. Explicit interests, language, distance preference, quiet hours, and discovery settings.
  3. Foreground, permissioned location capture, plus manual place selection when location is declined.
  4. Category-based nearby-place discovery using a licensed GIS/places provider.
  5. Personalized ranking and concise explanation of why an option is relevant.
  6. Tool-based AI reasoning over verified provider responses; no invented venues, schedules, availability, or route metrics.
  7. User-initiated recommendations and opt-in proactive suggestions for controlled triggers.
  8. Recommendation controls: save, dismiss, not interested, pause, and review learned preferences.
  9. A bounded memory system that distinguishes explicit facts from inferences.
  10. Basic general-wellbeing and nearby-care discovery, without prescriptions or unsupported clinical claims.
  11. Auditability for consent, tool invocations, source provenance, and consequential actions.

6. Flagship demo narratives

A. Cultural encounter

With a history interest and foreground location permission, ORVY discovers a nearby museum, verifies place details, and proposes a visit with a reason. Claims about live exhibitions must come from a current trustworthy source and otherwise be omitted.

B. Everyday fit

A user requests a quiet place to work or relax. ORVY considers preferences and verified nearby details; it explains why recommendations meet the request and can show directions.

C. Respecting control

A user revokes background discovery permission or deletes an inferred preference. Subsequent recommendations stop depending on revoked information, and the UI visibly confirms the change.

7. Explicit MVP exclusions

  • No full ORVY Core, Synapse, Reflex, decentralized consensus, blockchain or tokenization.
  • No banking transaction initiation, investment recommendations based on connected finance data, payment execution, or autonomous purchasing.
  • No medical diagnosis, prescribing, unverified medication dosing, or emergency-services replacement.
  • No unrestricted always-on tracking, secret profiling, or advertising-driven manipulation.
  • No multi-agent orchestration requirement; a single tool-using agent is sufficient.
  • No assumption that a PWA can perform reliable closed-app geofencing on both Android and iOS.

8. Human-first non-negotiables

  • Identity is not authority. Authentication establishes identity, not blanket permission to observe or act.
  • Memory is not authority. Past behavior never silently authorizes new collection, disclosure, or execution.
  • Recommendation is not execution. Consequential actions require a separate permission and confirmation.
  • Context is bounded. Retain only what a documented use case requires, with expiry and deletion behavior.
  • Evidence beats fluency. If a claim cannot be grounded, ORVY expresses uncertainty or does not state it.
  • User intent outranks engagement. Avoid notification overload and give meaningful opt-out controls.

9. Success criteria proposed for pilot testing

  • At least three reproducible end-to-end demonstration scenarios.
  • Zero unapproved external transactions or consequential actions.
  • Zero fabricated place names, operating hours, events, distances, or provider offers in the evaluation set.
  • At least 80% positive feedback on relevance among sampled pilot recommendations, measured transparently.
  • Consent withdrawal enforced before subsequent collection/processing.
  • A working installable web experience with clearly described background-location limitations.
  • Evidence of test coverage, source provenance, privacy controls, and cost/latency budgets.

Targets are project goals, not official contest judging criteria or guaranteed achievements.

10. Delivery stance

Documentation and human-readable HTML are prepared first in a separate GitHub draft area. Existing MHERC 2.0 and Codex work continue independently. The actual Companion application implementation starts only after a defined build gate. Initial deployment should reuse available Windows Dedicated Server, IIS, ASP.NET Core/C# MVC, SQL Server, and existing web-development tooling. OpenAI and mapping services remain usage-billed external dependencies.

11. Decision gates

  • G0 — Charter review: approve target user, core promise, MVP exclusions, and success criteria.
  • G1 — Architecture review: approve data boundaries, permissions, tool contracts, hosting, and threat model.
  • G2 — Implementation authorization: release a scoped Codex-ready build contract.
  • G3 — Pilot validation: demonstrate real places, evidence, consent revocation, and measured reliability.
  • G4 — Competition release: verify published entry requirements, presentation evidence, and demo readiness.

Current gate: G0, OPEN. Nothing in this document constitutes approval, freeze, release, or canonical promotion.

Human-readable derivative of companion/source/ORVY-COMP-CHARTER-v0.1.md. Source Markdown governs this draft; explicit review is required for promotion.

OverviewProduct CharterMVP ArchitectureDelivery Roadmap