Twenty relationship types
Parent, child, spouse, sibling, but also step-parents, adoption, extended kinship. Blended and adoptive families are not edge cases to work around.
Tunnda is a mobile app backed by a private infrastructure. Here is how the whole thing is built, and what that means in practice for your data.
Architecture
The order matters. The app does not fetch your data from a server: it holds your data, and backs it up.
Tunnda ships with a local SQLite database, not a cache. People, relationships, stories, recordings and media live there permanently. That is what makes the app fully usable without a network.
Your vault is backed up on our infrastructure, operated for us by Appwrite. That service exists only so that you can find your vault again. It is not used for analytics, for advertising, or for pooling between families.
The app integrates no audience measurement, behavioural tracking or third-party crash reporting tool. Technical error logs stay on your device, and you can switch them off.
Fingerprint, face and PIN code are verified by your phone. Those elements never pass through our servers.
How it works
The typical path of a family that is starting out.
The first account opens the family vault. Whoever opens it becomes its Elder: they are the one who will approve what goes in.
You add the people, the dates, the places, the photos. You record the voices of the elders. All of it happens offline, at the pace of visits and conversations.
You give each person a role: Co-admin, Contributor, Viewer, Youth. Everyone sees and does what matches their place.
Contributions go through the approval queue. The shared memory grows with what the Elder has recognised as accurate.
What Tunnda can model
Classic genealogy models break as soon as a family does not fit into a binary tree. Tunnda's model was designed for real families.
Parent, child, spouse, sibling, but also step-parents, adoption, extended kinship. Blended and adoptive families are not edge cases to work around.
Family memory is uncertain, and Tunnda accepts that. A piece of information can be qualified by how reliable it is, rather than presented as an established fact.
A profile distinguishes living people from people who have died. That is what makes it possible to anonymise the living automatically at export time.
An audio recording is a first-class object, attached to a person, with its type: name pronunciation, story, memory.