When you upload your blood work to GearCheck, what happens behind the scenes? The answer is a six-step pipeline that transforms a PDF, image, or manual entry into a comprehensive, personalized health analysis.
Think of it like a factory assembly line: each station adds a layer of processing, structuring, and intelligence. When the final product rolls off the line, you get a report that understands not just what your markers are, but what they mean for who you are.
Here is exactly what happens, from the moment you hit "upload" to the moment your report is ready.
Upload: Capture Your Data
You start by uploading your blood work. GearCheck accepts three input types: a PDF directly from the lab, a photo or screenshot of your printed results, or manual entry if you have the numbers on paper.
Direct lab PDFs. Text extracted via Fitz (MuPDF) in ~2 seconds.
JPEG, PNG from your phone. Processed via Gemini AI extraction.
During upload, GearCheck runs an immediate preflight check: it detects the file format, attempts text extraction, and runs a quick regex scan to count how many markers it can detect. If the panel looks too sparse — fewer than 10 markers found — it warns you before proceeding.
Estimated time: 2—3 seconds
Extraction: Structure the Unstructured
With the raw text in hand, the extraction engine goes to work. This is a hybrid system with two layers running in parallel for maximum coverage.
Over 200 curated regex patterns scan text for marker names, values, units, and reference ranges. Catches 85%+ of structured lab reports instantly.
For unusual names, non-standard formats, or image-based results, Gemini extracts structured data from unstructured text.
The result is a structured set of marker objects, each with a canonical name, value, unit, and reference range. Marker names are normalized to the GearCheck vocabulary so that "Testosteron," "Testosterone, Total," and "Total Testosterone" all map to the same canonical marker. This normalization is critical — without it, your results would be scattered across duplicates.
What about units? European labs report in mmol/L, US labs in mg/dL. The extraction engine converts everything to canonical units internally, so your report is consistent regardless of where you tested.
Estimated time: 30—60 seconds (bulk of pipeline time)
Reference Check: Compare Against the Right Baseline
Every extracted marker is compared against GearCheck's athletic-adjusted reference bands. These are not the generic lab ranges printed on your report. They are bands calibrated specifically for strength athletes and, where applicable, AAS users.
Best possible result for your profile
Outside range, needs review
Above critical threshold
Each marker receives a status label: OPTIMAL, NORMAL, ATTENTION, ACTION, or MONITOR. The reference bands adjust dynamically based on your user profile — sex, age, training status, and whether you have indicated AAS use. Markers that fall outside the athletic range are flagged for deeper analysis in the next step.
This step is where most generic lab analyzers stop. GearCheck keeps going.
Contextual Analysis: Interpret for Who You Are
This is where GearCheck separates from a generic lab analyzer. The Contextual Interpretation Engine applies 13+ rules that understand how AAS, training intensity, and body composition affect your markers. Each rule is a medical logic pattern derived from sports medicine literature.
Convert false alarms to contextual flags. Normal lab says ACTION but we know it is expected — e.g., elevated AST from training.
Catch what standard ranges miss — e.g., normal creatinine with elevated Cystatin C flags hidden kidney stress.
Here are some of the key patterns the engine detects:
- High testosterone + normal liver markers = intended effect, not pathology
- Low eGFR + high creatinine + normal Cystatin C = muscle mass, not kidney disease
- Elevated AST/ALT + elevated CK + normal GGT = muscle leak, not hepatotoxicity
- Low HDL + high ApoB = AAS-typical pattern, not genetic dyslipidemia
- Elevated hematocrit + normal platelets + no symptoms = expected, proceed with monitoring
Each rule can downgrade a marker's priority from ACTION to CONTEXTUAL, or upgrade it from NORMAL to ATTENTION, depending on the full picture. Context can flip a marker from alarming to expected in milliseconds.
Assembly: Build the Big Picture
With all markers analyzed and contextualized, the assembly engine builds your complete report. This is where individual data points become a coherent story.
Weighted composite of all marker statuses, 0—100. Contextual flags get adjusted weights so the score reflects real risk, not pharmacological noise.
Top 3—5 markers needing attention, ranked by priority and severity with actionable guidance.
Connected marker patterns that tell a story: "High T → suppressed SHBG → low HDL → elevated LDL."
Narrative written in natural language explaining your overall health picture, written for your level of understanding.
The assembly engine also identifies organ system clusters: if multiple liver markers are flagged together, the penalty multiplies because concurrent signals are more significant than isolated ones. The result is a risk assessment that matches how real physiology works.
Output: Deliver Actionable Intelligence
The final output is a complete, readable report with actionable guidance — not just numbers, but what to do about them.
Complete analysis in the GearCheck dashboard, organized by marker category with status indicators.
Personalized recommendations: which markers to retest, when, and what lifestyle or protocol changes to consider.
When to draw blood next, based on your current status and risk profile. No guesswork.
Download a formatted PDF for your doctor or personal records. Professional, clean, shareable.
Every report also includes a comparison to your previous results, so you can see trends at a glance. A single report tells you where you are. A series of reports tells you where you are going.
All in Under 5 Minutes
⚙️Most lab analyzers stop at Step 3: here are your markers, here are the ranges. GearCheck adds Steps 4 through 6 — the contextual interpretation that turns data into a decision-making tool.
— GearCheck Engineering
Personalized
Not generic ranges — ranges calibrated for strength athletes and AAS users specifically.
Contextual
13+ rules that understand pharmacology, training effects, and body composition.
Actionable
Not just what is wrong — what to do about it, when to retest, and how to adjust.