Skip to Content

Graph

The knowledge graph is the visual investigation tool at the centre of a case. It draws your assets and findings as connected nodes, with the relationships between them as links, so you can see how things relate instead of reading down a table. It’s where a tangle of findings becomes a picture you can reason about.


What’s on the graph

NodeRepresents
AssetA document, file, record, or other item from a source — shown with its type icon
FindingA signal a detector raised — shown as a coloured dot, where the colour is its severity

Nodes are joined by relationships — “contains,” “references,” “owns,” “accessed,” “sent to,” and more. Each relationship comes from one of three places, so you always know how sure to be of it:

Where it came fromMeaning
From the sourceDiscovered in the source itself during scanning (e.g. a page hierarchy)
InferredWorked out by Classifyre from things the items share (e.g. the same value in two places)
ManualDrawn by you, to capture a connection you know about

Exploring

The graph has three simple modes for working with it:

ModeWhat it does
ExploreClick a node to inspect it, drag to rearrange, and pan around the canvas.
ConnectClick one node, then another, to draw a relationship between them.
Find pathPick two nodes and the graph highlights the shortest chain of relationships linking them.

You can also expand the graph outward from any node to pull in its related items, growing the picture as your investigation widens.


Visual cues

The graph encodes a lot at a glance:

  • Evidence ring — items attached as evidence are circled, so case evidence stands out from surrounding context.
  • Hypothesis dots — small coloured marks show which hypotheses a node is linked to, in each hypothesis’s colour.
  • Cross-hypothesis marker — a node tied to more than one hypothesis is flagged, so contested evidence is easy to spot.
  • Grouped findings — an asset with many findings shows a compact “+N” badge you can expand when you want the detail.
  • Highlighted path — the chain between two nodes you’re tracing is picked out clearly.

Asking questions of the graph

Rather than clicking around manually, you can ask the graph focused questions that expand it along meaningful relationships:

QuestionShows you
Who touched this?Who or what accessed, read, or ran the item
Where did it come from?The upstream lineage that produced it
Where did it go?The downstream lineage it fed into
Who has access?Ownership and access relationships
Email trailWho it was sent to or mentioned with
Similar findingsOther items carrying the same kind of signal

These turn the graph into an investigative interview: ask a question, and the relevant relationships light up.


Working directly on the graph

Right-click any node or relationship to act on it without leaving the graph — add or remove evidence, attach a finding, link a node to a hypothesis (for or against), draw or trace a connection, expand to related items, or open the item’s full detail. Relationships you drew yourself can be renamed or deleted; relationships that came from the source or were inferred are read-only, so the underlying truth isn’t accidentally rewritten.


Side panels

A few panels keep the big picture in view while you work:

PanelHelps you
HypothesesSee every hypothesis with its verdict, confidence, and for/against count — and spotlight the evidence linked to one
Highlight filtersDim everything except the sources or detectors you care about
Relationship filtersShow or hide specific kinds of connection
Legend & statsA key to the shapes and colours, plus live counts of assets, findings, evidence, and links

Together they let you move from a wide view of the whole case down to the single relationship that cracks it.

Last updated on