Independent AI-Assisted Courtroom Intelligence

Evidence, explained in real time.

A continuously updated analytical layer over public courtroom testimony—mapping competing theories, surfacing contradictions, measuring evidentiary strength, and preserving a transparent audit trail for every model movement.

Illustrative interface. Scores shown here are placeholders, not factual findings, diagnoses, legal conclusions, or accusations. Production analysis should link every material assertion and score change to primary-source testimony or court records.

Case theories

Primary legal positions and a separately tracked public hypothesis

Updated 10:42 AM

Prosecution Theory

Criminally Responsible

▲ 4.0 today

Evidence supporting the proposition that Lindsay caused the deaths while possessing the legally required criminal responsibility.

61/100 evidence support
Defense Theory

Not Criminally Responsible

▼ 2.0 today

Evidence supporting the proposition that mental disease or defect prevented legal criminal responsibility at the relevant time.

67/100 evidence support
Independent Public Hypothesis

Patrick-as-Perpetrator Theory

Tracks evidence that supports or undermines the theory that Patrick Clancy was the physical perpetrator. This hypothesis is distinct from the prosecution and defense positions being litigated. The score represents evidentiary support for the theory—not a probability of guilt or factual accusation.

16/100 EVIDENCE SUPPORT
1,842Transcript segments analyzed
126Evidence items indexed
34Contradictions flagged
98.7%Source-linked model actions

Evidence intelligence

Independent dimensions of the record

Scores do not need to total 100

Psychosis Evidence

Mental state
58

Strength of evidence consistent with psychosis or severe mental disturbance.

Planning / Premeditation

Intent
71

Evidence of purposeful sequencing, preparation, planning, or deliberate conduct.

Timeline Consistency

Chronology
63

Alignment among testimony, communications, timestamps, and physical evidence.

Prosecution Burden Strength

Legal burden
69

How strongly the current record supports the prosecution meeting its burden.

Unresolved Evidence

Open issues
44

Material contradictions, disputed facts, and evidentiary gaps not yet reconciled.

Source Coverage

Data quality
82

How much of the current analysis is directly connected to source transcript material.

Live intelligence feed

What changed, why it matters, and which score moved

Audit-first design

10:41 AMPsychiatric testimonyHigh relevanceSource-linked

New mental-state evidence entered the model

The witness described observations relevant to Lindsay’s mental condition near the relevant time. The system cross-referenced the testimony against prior psychiatric evidence and marked two points for expert-conflict review.

Psychosis -3.1Criminal Responsibility +2.4Defense Theory -2.0
10:28 AMCross examinationConflict detected

Prior characterization challenged

A statement appears inconsistent with another description already in the record. The system preserved both excerpts and increased the unresolved-evidence score rather than selecting a side automatically.

Unresolved +4.2Burden Strength -1.1
10:11 AMTimelineCorroborated

Time reference reconciles with an existing timestamp

A newly stated time point appears consistent with an already indexed communication event, increasing confidence in the chronology.

Timeline +3.6Planning +1.5

Recent model actions

Transparent, inspectable decision log

Every movement should be reproducible

Time
Evidence event
Classification
Confidence
Source
10:41:22
Psychiatric observation added
Mental state
0.92
Transcript #1842
10:28:09
Statement inconsistency detected
Contradiction
0.88
#1828 ↔ #1044
10:11:47
Timeline event corroborated
Chronology
0.95
Transcript #1802

Automation architecture

The technology behind the public interface

Designed for continuous operation

01 — INGESTCapture new transcript segments with speaker and timestamp metadata.
02 — NORMALIZEClean text, resolve speakers, deduplicate repeated caption fragments.
03 — ANALYZEClassify evidence, detect contradictions, and map testimony to legal questions.
04 — SCOREApply a fixed evidence rubric with bounded changes and confidence controls.
05 — PUBLISHUpdate dashboard state, explanation cards, and the permanent audit trail.