Cortex Grid

What is Cortex Grid?

Investigative work fails when a plausible narrative outruns its evidence. Cortex Grid is built to prevent that: evidence is hashed and logged on intake so its handling can be demonstrated, entities are resolved across sources into a reviewable graph, media is assessed for authenticity and Content Credentials, and threat is scored on a component-by-component rubric. Most importantly, what is known, what is inferred and what is alleged are kept visibly distinct at every point — the discipline that separates an investigation from a story.

Problems it solves

• Preserve evidence with demonstrable chain of custody
• Resolve entities across fragmented sources
• Build a verifiable timeline from digital evidence
• Assess media authenticity and Content Credentials
• Score threat on a component-by-component rubric
• Keep fact, inference and allegation visibly separate

Key capabilities

Chain of Custody

Hash and log evidence on intake so its handling can be demonstrated later.

Entity Resolution

Connect people, accounts and artefacts across sources into one reviewable graph.

Timeline Construction

Assemble dated events into a timeline, each entry linked to its source.

Media Authenticity

Assess whether media shows signs of manipulation, and report Content Credentials where present.

Fact / Inference Separation

Keep what is known, what is inferred and what is alleged visibly distinct.

Risk Scoring

Score a change, a lot or a habitat on a published rubric you can read component by component.

Evidence Trail

Every figure carries the record it came from, so a reader can check the answer instead of trusting it.

How it is built

Hashed intake

Evidence is hashed and logged on arrival so handling can be demonstrated later.

Entity graph

People, accounts and artefacts resolved across sources into one reviewable structure.

Authenticity assessment

Media examined for manipulation signals, with Content Credentials reported where present.

Epistemic status everywhere

Fact, inference and allegation are distinct in every view.

How it works

How the work is structured, and what each step's output rests on.

1. 1

   Preserve on intake

   Source

   Hash and log evidence with its provenance.

2. 2

   Resolve entities

   AI-inferred

   Connect people, accounts and artefacts across sources.

3. 3

   Assess media

   Observed

   Examine authenticity and report Content Credentials.

4. 4

   Build the timeline

   Observed

   Assemble dated events with their supporting evidence.

5. 5

   Score threat

   AI-inferred

   Apply the rubric component by component.

6. 6

   Separate status

   Human-verified

   Mark each element as fact, inference or allegation.

Example use cases

Each example has its own page you can read and share.

Ask Cortex Grid

Illustrative examples. Results depend on your own data.

• “Build a timeline from this evidence set.”
• “How are these accounts connected?”
• “Does this media show signs of manipulation?”
• “Which parts of this case file are inference rather than fact?”

Security, governance and human control

Where the software deliberately limits itself.

Questions

Does it decide guilt or attribution?

No. It organises evidence, resolves entities and scores threat on a published rubric. Conclusions are for investigators, and the tool keeps inference labelled as inference.

How is chain of custody handled?

Evidence is hashed and logged at intake, and handling is recorded so it can be demonstrated.