← Back to Blog
Roadmap

Roadmap Preview: GeraMind v0 to v1 — Phases, Milestones, Open Questions

Published 21 April 2026 · 4 min read

Coming soon — join the waitlist

Quick answer. Three phases: v0.1 (vault storage + scope vocabulary + token format, Q4 2026), v0.5 (minimisation engine + audit dashboard + 3 ingest adapters, mid-2027), v1.0 (multi-vault federation + regulator-format audit export + enterprise onboarding, late 2027). Dates are targets, not commitments. Open questions listed per phase.

Phase v0.1 — storage and tokens (Q4 2026)

  • At-rest encryption with per-scope key hierarchy.
  • Scope vocabulary v1 (published and versioned).
  • Purpose-bound token format (JWS, narrow scope, short expiry).
  • Basic query surface for three scope families: preferences, health, documents.
  • Reference client library: TypeScript + Python + Swift.
  • First production integrations: GeraClinic, GeraEats.

Open question that could push this back: the scope vocabulary. We are converging with Solid ontologies and Apple’s taxonomies, but convergence is slow.

Phase v0.5 — minimisation and audit (mid-2027)

  • Minimisation engine as the default query path.
  • Audit dashboard for end users: inspect, revoke, export.
  • Weekly digest emails with anomaly flagging.
  • Ingest adapters: Apple Health, Gmail, Google Fit.
  • Per-app per-scope query budgets.
  • Integration with GeraNexus consent flow.

Open questions: cross-scope inference-attack mitigations (empirical, not closed-form), and the UX for “raw content released” flagging so it is informative without being alarmist.

Phase v1.0 — federation and enterprise (late 2027)

  • Multi-vault federation: storage can be split.
  • Audit export in regulator-readable format (ICO, DPA, ANPD, ePrivacy-compatible).
  • Self-hosted vault image: Docker + systemd unit + a healthchecks.io template.
  • Enterprise onboarding path with BAA-equivalent for health scopes.
  • v1.0 spec freeze with a 2-year stability commitment.

Open question: regulator acceptance of the audit format. Conversations are early. If we cannot get at least one regulator to accept the format as evidence, v1 ships the format and enterprise adoption carries the convergence pressure.

What we will not do in v1

  • LLM fine-tuning on vault data. Ever.
  • Cross-user analytics. Ever.
  • Ads on top of vault data. Ever.
  • Payment rail hosting. That is GeraCash or Stripe.

Scope discipline ships specs. Every nice-to-have we deferred to v2 is a nice-to-have that did not block v1.

How to follow or contribute

Draft spec: geramind.com/spec. GitHub issues are open. If you work on Solid, Inrupt, Apple data APIs, DTP, or a similar project, we want early eyes on v0.1 before freeze. Email [email protected] or join the waitlist below.

Help us design the vault.

Join the waitlist