FirstDomino

How it works.

Records go in, a graph connects them, and FirstDomino asks for the missing fact. Possible patterns for a clinician to review; never a conclusion, never a medicine change.

  1. 1

    Records

    Letters, pharmacy records, the caregiver. Each fact keeps its source, day and certainty.

  2. 2

    Graph

    Every medicine, symptom, record and finding is a node in FalkorDB.

  3. 3

    Questions

    It asks the person who knows for the fact that settles most. The clinician decides.

The graph model

A finding is its own node, linked to its facts, its pattern and the fact it still needs. So one changed fact moves exactly the right findings.

hashashasfromfirstmatchescitestoneedsasks forofonPatientMedicine 1SymptomMedicine 2RecordPattern (cited)FindingMissing factQuestionVersionDecision
Simplified. Every walk starts from one patient, so one patient’s facts never mix with another’s.

What FalkorDB does here

Finds the chain
One walk from the patient: medicine → class → pattern → symptom → second medicine.
Knows what depends on what
A changed fact points at the findings that rest on it. Only those are re-checked.
Picks the next question
It asks for the missing fact the most open findings wait on.
Keeps every version
Old values, finding revisions and decisions stay linked to what they were about.
Refuses stale writes
Each write checks the case version in one query, so no one overwrites a newer case.

In any review, open “How the graph found this” and turn on “Show technical details” to see the named queries and their Cypher.

AI assistants can use the same review through FirstDomino’s read-only MCP server, with the evidence IDs behind every answer.

Take the 2-minute guided tourExplore the stories