Current Status

This page is generated from Marvin's shared status model. Run npm run marvin:render-status-doc to refresh it manually.

As of Thursday, July 30, 2026, Project Marvin is a serious local proof-of-product, not a fully proven production-finished product.

Current coverage

The repo currently supports all of the following locally:

• bidirectional route planning across Microsoft 365, Outlook, Google, and optional Apple / CalDAV calendars
• private-by-default mirrored events
• family-calendar detail overrides
• automatic per-source prefixes such as WORK: and FAMILY:
• source-timezone preservation in mirror payloads
• Marvin-owned Microsoft and Google OAuth start/callback flow
• Marvin-owned Apple / CalDAV credential validation
• Marvin account creation, staged onboarding, sign-in gating, and ongoing calendar management UI
• scriptable local install, bootstrap, verification, and Azure deployment-plan generation
• generated doctor/status reporting tied to the shared requirement model

Requirement matrix

Requirement	Repo truth on 2026-07-30	Strongest local evidence	Remaining gap
An event created or accepted in any connected calendar mirrors to every other connected calendar.	Partially proven locally	npm run marvin:smoke-live, npm run marvin:smoke-live-readiness, npm run marvin:smoke-subscriptions, npm run marvin:smoke-runtime-webhook-wake	Not yet proven against real customer-owned live calendars.
Marvin mirrors that event to every other connected calendar.	Partially proven locally	npm run marvin:smoke-live, npm run marvin:smoke-live-readiness, npm run marvin:smoke-subscriptions, npm run marvin:smoke-runtime-webhook-wake	Not yet proven in a real always-on deployed runtime.
Mirrored events are private by default.	Partially proven locally	npm run marvin:smoke-live	Not yet proven against real tenant data and viewer permissions.
Selected target calendars, such as family calendars, can receive full detail instead.	Partially proven locally	npm run marvin:smoke-live, npm run marvin:smoke-account-management	Not yet proven in real customer calendars.
Every mirrored event carries the source calendar prefix.	Proven locally	npm run marvin:smoke-live, npm run marvin:smoke-account-management	Not yet proven end-to-end in real tenant-backed writes.
Timezone behavior follows the source event instead of being hardcoded.	Proven locally	npm run marvin:smoke-live, npm run marvin:smoke-microsoft-timezone	Not yet proven with real travel and provider edge cases.
Real provider authentication and connected-account validation must work before Marvin treats a calendar as ready.	Partially proven locally	npm run marvin:smoke-onboard-api, npm run marvin:smoke-operator-journey, npm run marvin:smoke-connection-validation, npm run marvin:smoke-auth-gating	Still not proven across real hosted runtimes and real customer-owned provider tenants.
Simple onboarding and ongoing account-management UI must exist for Marvin operators.	Partially proven locally	npm run marvin:smoke-ui-surface, npm run marvin:smoke-operator-journey, npm run marvin:smoke-onboarding-guidance	Still not proven as a finished production UX across real tenants and hosted runtime.
Ongoing management must allow adding, removing, and updating calendars and mirror policy.	Proven locally	npm run marvin:smoke-account-management, npm run marvin:smoke-ui-surface	Not yet proven in a real deployed multi-user environment.
Installer, bootstrap, verification, and deployment flows must be scriptable.	Partially proven locally	npm run marvin:smoke-install, npm run marvin:smoke-bootstrap, npm run marvin:smoke-docs-commands, npm run marvin:smoke-deploy-plan	Full production deployment and operations proof is still missing.
Microsoft 365, Outlook, and Google are in scope. Apple / CalDAV remains optional.	Partially proven locally	npm run marvin:smoke-live, npm run marvin:smoke-operator-journey, npm run marvin:smoke-caldav-live, npm run marvin:smoke-outlook	Real end-to-end customer proof across all providers is still missing.
Repository documentation must reflect the true architecture, implementation status, setup, deployment, and testing process.	Partially proven locally	npm run marvin:smoke-docs-commands, npm run marvin:smoke-status-reporting, npm run docs:build	Docs render and status wiring are proven locally, but they still describe a locally proven product rather than a fully completed production deployment.

Coverage summary

• Total requirements tracked: 12
• Proven locally: 3
• Partially proven locally: 9
• Missing: 0

Biggest remaining gaps

Marvin still does not have strong proof of all of the following:

• real customer-owned live calendars syncing end to end across real tenants
• production-grade hosted secret handling and operational hardening
• fully proven always-on hosted runtime lifecycle
• fully zero-touch provider-app creation across every Microsoft and Google tenant
• a final documentation set that proves completed production deployment rather than a strong local proof

Use the repo to verify status

If you want the repo to tell you what is true right now, use:

npm run marvin:doctor

For the full JSON report, including requirement coverage and next gaps:

node scripts/marvin-doctor.mjs --json

For the fuller written requirement audit, read:

• Requirements
• Getting Started
• Marvin Engine