Four solutions.
One verification layer.
One layer. Every calibration defensible.
WLTR turns raw instrument output into a verified, traceable, QA-approved calibration record. No retyping. No hidden math. No spreadsheet archaeology when an inspector asks for proof.
Deterministic
The same source data produces the same result, every time.
Traceable
Every value links back to immutable raw instrument text.
Defensible
Readiness, exclusions, criteria and QA approval stay with the record.
Raw output becomes an auditable decision.
Designed around regulated laboratory reality
EPA 8260
ISO/IEC 17025
GLP
GMP
ALCOA+
Four solutions. One verification layer.
Each removes a specific point of failure between the instrument and the signed record.
01
Data Integrity
Raw instrument text parsed, hashed and preserved — never retyped.
Raw instrument text is parsed, hashed and preserved — every number traces to its source.
02
Calibration Intelligence
Testing models and weighting schemes one at a time hides the best fit.
Every candidate model scored against 19 acceptance criteria, with the simplest passing model recommended.
03
Multi-Analyte Visibility
Whole-group pass/fail triage — reviewers start at the exceptions, not page one.
04
Audit Defensibility
Inspections surface decisions nobody documented at the time.
Frozen configs, owned exclusions, named QA approval — evidence stays with the record.
Your instrument builds a curve. It does not prove the curve is right.
Analysts still export results, rebuild calculations in Excel, test weighting schemes by trial and error, hand-check acceptance criteria, and explain undocumented decisions months later. WLTR replaces that vulnerable middle layer.
Instrument software + spreadsheet review
- Raw values are copied, pasted or retyped into uncontrolled worksheets.
- Models and weighting schemes are tested one at a time.
- ICV verification and %Diff checks live in separate files.
- Exclusions can disappear without a durable reason or owner.
calibration intelligence
- Raw instrument text is parsed, hashed and preserved as the source record.
- Every candidate model is evaluated against 19 acceptance criteria for each analyte.
- Acceptance criteria and ICV performance are verified in one system.
- Every exclusion, configuration and approval remains attached to the record.
A cleaner architecture for calibration oversight.
The design is deliberately restrained: source integrity first, computation second, visibility third, approval last. Humans remain responsible for decisions. The system makes those decisions inspectable.
The record starts at the raw text — and is never retyped.
Ingest instrument output directly, parse it deterministically, and store the original text immutably with a SHA-256 hash so every downstream value traces to its source.
- Raw instrument text parsed and hashed at upload
- Run-type aware — an ICV can never silently enter a curve
- An immutable source record behind every reported number
app.wltr.io / upload-run
Every candidate model. Every analyte. One view.
Evaluate every candidate model for every analyte against 19 acceptance criteria and display the calibration with simultaneous overlays — point-level %Diff, model fit and ICV recovery — without rebuilding the curve by hand.
- All candidate models computed simultaneously, server-side
- Every level scored against the active threshold
- The simplest passing model is recommended
app.wltr.io / calibration-intelligence
Whole-group visibility. Nothing slips through.
Aggregate calibration curves across every analyte and surface per-analyte pass/fail states, spotlighting trends and anomalies. Reviewers start with the exceptions instead of hunting for them.
- Whole-group computation across all analytes
- Per-analyte pass/fail triage on one screen
- Aliases resolved once, matched automatically after
app.wltr.io / analyte-mapping
Audit-ready output for any inspection.
Freeze the method configuration at compute time, log exclusions with user and reason, verify the ICV, and route the completed record through a named QA approval gate.
- Readiness validations gate every compute
- Method config frozen — historical results never move
- Role-separated QA approval with a named signature
app.wltr.io / calibration-groups
From source text to signed record.
Seven steps replace the fragmented spreadsheet stage without pretending laboratory judgment can be automated out of existence.
Configure the laboratory
Define instruments, methods, acceptance criteria, users and QA roles.
Upload calibration and ICV runs
Capture raw instrument text, metadata and source hashes before analysis begins.
Resolve analyte mappings
Match raw compound names to canonical analytes and preserve aliases for future runs.
Build the calibration group
Compute all candidate models
Generate the candidate models for each analyte and evaluate every level against the active criteria.
Generate the evidence record
Package selected models, exclusions, ICV performance and method context into a consistent report.
Route to QA approval
Complete role-separated review with a named approver and permanent audit timeline.
Manual review has a documented failure rate. That is the point.
Review by exception
Whole-group computation makes failed analytes visible immediately instead of burying them in repetitive manual review.
Criteria frozen in context
The active method configuration is snapshotted at compute time, so later changes do not rewrite historical results.
Every exclusion has an owner
Removing a point becomes a recorded decision with who, when and why, not a disappearing spreadsheet row.