User Manual

Everything you need to get started with TRACER and maximize the quality of your analysis results.

Part 1 — How to Use TRACER

A step-by-step guide to running a deception analysis from start to finish.

1Create an Account

Sign up at the login page with your email or Google account. Once registered, choose a subscription plan that fits your needs. Plans range from a single scan to professional packages with hundreds of analyses per month.

2Start a New Case

From your dashboard, click New Case. You will be guided through a step-by-step wizard. Begin by entering the Subject's name, Case ID, and your Investigator name.

3Define the Issue Under Investigation (IUI)

Enter the specific issue you are investigating. The IUI is the central focus of the analysis — TRACER evaluates the Subject's language in relation to this issue. You can define multiple IUIs if the case involves more than one matter.

Single IUI example:"Did the Subject steal funds from the company account between January and March?"

Multiple IUI example:(1) "Did the Subject steal funds from the company account?" and (2) "Did the Subject access restricted financial records without authorization?" — each IUI is evaluated independently, allowing TRACER to isolate deception indicators relevant to each specific allegation.

4Enter Known Truths, Known Lies, and Evidence

Known Truths— Statements you can confirm are 100% true. These anchor TRACER's accuracy threshold and calibrate the analysis against verified facts. For example, if surveillance footage confirms the Subject was at a specific location at 9:00 AM, entering that confirmed fact as a Known Truth allows TRACER to evaluate the Subject's statements about that time period against a verified baseline.

Known Lies — Statements you can confirm are 100% false. For example, if the Subject claimed they were home at 6:00 AM but later admitted to being at the scene, that prior denial can be entered as a Known Lie. Note: Known Lies should reflect intentional falsehoods — if a misstatement may have been unintentional or a memory error, enter it as Evidence rather than a Known Lie.

Evidence — Relevant information that is not yet confirmed. This is used as context but does not directly affect scoring.

If the same item appears in both Known Truths and Known Lies, the wizard will prompt you to resolve the conflict before proceeding.

5Upload the Transcript

Upload the Subject's statement as a text file, PDF, or Word document. TRACER accepts both Q&A format (interviewer questions followed by Subject answers) and single narrative (continuous statement). Transcripts should be as verbatim as possible — include hesitations, partial words, and fillers for best results.

Important: PDF files must contain readable, selectable text. PDFs created by scanning a physical document (image-based PDFs) may not extract correctly. If your PDF is scanned, paste the transcript text directly into the text box instead.

6Provide Baseline Data (Recommended)

Baseline data helps TRACER distinguish the Subject's natural speech patterns from potential deception indicators. Enter baseline Q&A pairs— questions and answers where the Subject's responses are known to be truthful.

You can also enter uncertainty phrases— phrases the Subject naturally uses when genuinely unsure (e.g., "I'm not sure," "I think"). These prevent natural hedging from being misclassified.

Without baseline data, all indicators are scored at full value with no exemptions for natural speech patterns.

7Run the Analysis

Click Run Analysis. TRACER's pipeline will process the transcript through multiple analysis stages. This typically takes 3–7 minutes depending on statement length. You can monitor progress in real time.

8Review the Report

Once complete, your report is available immediately. The report includes:

  • Executive Summary with per-IUI findings
  • Per-response narrative analysis
  • Timeline of events with gap detection
  • Evidence conflict analysis
  • Follow-up questions for subsequent interviews
  • Pattern recognition insights
  • Heat map summary of all responses
  • Hypothetical narrative reconstruction

You can view the report online or export it as a professional PDF.

Understanding Assessment Bands

Each response in the report receives one of four assessment classifications:

Likely Truthful

Strongly supported by consistency

Inconclusive

Not enough data to assess

Sensitive

Area where Subject may be hiding details

Likely Deceptive

Strong deception indicators present

Part 2 — Good Questioning Techniques

How to maximize TRACER's accuracy and reliability through better interviewing practices.

Building a Strong Baseline

Start every interview with 5–10 minutes of baseline questions. Ask irrelevant, low-stakes factual questions that encourage truthful answers:

  • "What's your favorite food?"
  • "Where did you grow up?"
  • "What do you do for work?"
  • "Walk me through your typical morning routine."

This captures the Subject's normal truthful style — pronoun usage, hedging patterns, filler words, and fluency. Without baseline data, TRACER cannot distinguish natural speech quirks from deception indicators, resulting in potential false positives.

Do not skip this step. It is the single most important factor in analysis accuracy.

Effective Question Design

Direct, Neutral, Open Questions

Encourage the Subject to provide uninterrupted narratives. Open-ended questions produce the richest data for analysis.

"Walk me through what happened that morning."

Specific, Anchored Follow-ups

When following up, anchor your question to the Subject's own words. This creates accountability without leading.

"You said you left the house. What made you leave then?"

Balanced Pacing

Allow the Subject to complete their full answer before asking the next question. Cutting someone off mid-sentence degrades the data TRACER can analyze and may create false gaps in the narrative.

Neutral Tone

Questions should be fact-seeking, not accusatory — unless you are deliberately testing a specific point with preserved evidence.

Practices to Avoid

The following significantly degrade TRACER's ability to assess truthfulness:

Swearing at or belittling the Subject

Forces TRACER to suppress most analysis rules, leaving only minimal detection active.

Cutting the Subject off mid-sentence

Creates incomplete data that may produce false gaps or insufficient scoring.

Talking over the Subject

Contaminates the transcript and makes it impossible to attribute statements accurately.

Coercive framing

Phrases like "You'd better tell me the truth" create stress responses that mask genuine deception indicators.

Overloading with compound questions

Subjects may only answer one fragment, leaving other parts unaddressed and unanalyzable.

Leading questions

Risk contaminating the Subject's answer with the Interviewer's own wording.

Yes/No questions without follow-up

Shut down narrative and provide minimal data for analysis. Pair with open-ended follow-ups.

Rule of thumb: Any prompt that degrades rapport degrades TRACER's analysis.

Clarification Strategy

After analysis, TRACER generates targeted follow-up questions for you to use in subsequent interviews. These follow-ups are anchored to the Subject's own words and designed to resolve gaps, clarify anomalies, and test unresolved indicators.

When following up, always re-anchor to the Subject's own phrasing. Do not fabricate or assume missing details — TRACER only analyzes what was actually said.

Example: Good vs. Bad Questioning

Bad Question

"Why don't you admit you did it? Stop lying."

Result: Classified as belligerent. Suppresses most analysis rules. The Subject's answer becomes largely unusable.

Good Question

"Walk me through what happened after you left the house that evening."

Result: Fully analyzable. TRACER can detect time gaps, conflict cues, speech pace anomalies, and potential admissions or denials.

Important Reminder

TRACER evaluates language patterns only. It does not determine guilt, innocence, or legal culpability. All outputs are linguistic likelihoods and anomalies intended to support — not replace — professional judgment. TRACER is a decision-support tool for trained investigators and should never be used as the sole basis for legal, employment, or personal decisions.