Aerial: connected records, explicit permissions

From Family Shapes to Aerial

I started with Family Shapes to make family and donor relationships visible. Aerial carries that work into a larger question: how do families, donors, and institutions keep the same records connected over time, with clear responsibility for updates? I now position it as a trust layer for the donor-gamete industry. The two Family Shapes projects remain historical records, including their original screenshots; this is the current project.

One connected story across three portals

I built a fictional cast for the product demos so the same people and records can be followed across roles. Maya Brooks sees her family and sibling network, Evan Park maintains donor information, and Dr. Lena Ortiz works in Harbor Cryobank. The current family dashboard distinguishes 23 recorded families and 34 children from the 12 families whose identities are visible. Those numbers describe synthetic demonstration data, not customer adoption or industry coverage.

Transparency has to preserve choice

I chose to separate recorded totals from permission to see a person. Identity, photos, location, and contact details have separate sibling and donor audiences; institutional records and care access follow their own rules. The family canvas makes biological, legal-parent, and donor relationships visible without treating those roles as interchangeable. The screenshots show saved sharing choices and the connected canvas, while the schema and permission tests document the underlying constraints.

Records need a human follow-through

I expanded the workflow from viewing a network to handling capacity, medical notices, correction requests, and source discrepancies. Institutions can track recorded families separately from reservations, assign follow-up, and review versioned evidence. AI comparison is useful for finding disagreements in authorized documents, but a human reviewer must inspect the sources and record a decision and rationale. A matching name or an AI finding is not proof that two donor identities are the same.

The design and demo work are part of the evidence

I used Codex for implementation, synthetic fixture preparation, visual iteration, and repeatable demo capture. I reviewed the resulting pages and videos, revised the narrative, and kept the connected-atlas direction across the marketing site. The September 12–13 releases brought the lavender visual system and a Resources hub with a glossary, industry timeline, and videos that play inline or in a modal. The artifacts here connect that design work to the product instead of treating a polished homepage as evidence that every workflow is complete.

What is working, and what remains open

Source baseline: aerial-platform main 734411bb, September 14, 2026. Current application screenshots use the fictional YC cast. Source implementation evidence: feature matrix and schema.

Dataset Preview

• Raw CSV row/table content is available in the source artifact.