Getting Started
From a fresh install to your first answer is three steps.
1. Open your GDAT database
- At the top of the window, click … and pick your GDAT
.db file
(or paste its path into the “path to GDAT .db” box).
- Click Open. The status line shows something like
“1,234 persons · read-only · map 5,678 tokens” — DGAI has indexed
your data and minted a token for every person, surname, place, and group.
2. Choose an AI
- Pick a provider from the backend dropdown (e.g. Anthropic for cloud,
or Ollama for a local AI on your own machine).
- For a cloud provider, open Privacy & Debug → Settings and paste your
API key (or set the provider's environment variable). Then press Open
again so the key takes effect.
- Optionally pick a specific model; click ↻ to refresh the list.
See Choosing an AI Backend for the full list and trade-offs.
3. Ask a question
- Type in the box at the bottom (“ask about your matches…”) and press
Enter or click Send.
- Watch the privacy badge near the top: Pseudonymized (green) means
every identity is tokenized before sending.
- Read the answer — names shown in
green highlight were restored locally from tokens. Hover any reply to see exactly which tokens
were sent.
📷 Screenshot: The DGAI chat window — database path, backend/model pickers, privacy badge, and a first answer.
A good first question: “Give me a summary of this database, then my
top 5 matches and which ancestral line they cluster on.” DGAI will ground itself on your data and
walk the matches.
If you typed a relative's name, DGAI shows a note like
“sent as: ‘Mary Jones’ → [M1234]” — confirmation that the name
was swapped for a token before the message left your machine.