Analysis Reference

Complete reference for every metric in the SessionAnalysis object returned by getAnalysis(). See Analysis for usage examples and indicator descriptions.

SessionAnalysis

Top-level fields on the analysis result object.

Metric	Type	Description
`version`	`string`	Schema version of the analysis output (e.g. “1.0.0”).
`locale`	`string`	Browser locale at time of analysis (e.g. “en-US”). Empty in Node.js.
`analyzedAt`	`number`	Unix timestamp (ms) when the analysis was performed.
`sufficientData`	`boolean`	Whether enough data was captured for meaningful analysis (30+ keydowns, 5+ seconds).
`keydownCount`	`number`	Total keydown events recorded during the session.
`integrity`	`IntegrityProof`	SHA-256 hash chain over raw events for tamper detection.
`sessionTimeline`	`SessionSegment[]`	Classified timeline segments (typing, paste, pause, tabAway, autocomplete, navigating).
`initialTextLength`	`number`	Character count of any pre-existing text when tracking started.
`initialTextHash`	`string`	SHA-256 hash of the initial text content.
`finalTextLength`	`number`	Character count of the text when tracking stopped.
`finalTextHash`	`string`	SHA-256 hash of the final text content.
`activeTime`	`ActiveTimeResult`	Active vs idle time breakdown for the entire session.
`outputSignature`	`string`	Hex-encoded DER signature over the analysis output for tamper verification.
`signedPayload`	`string`	The exact JSON bytes that were signed, for cross-language verification. (optional)

Content Origin

Access via analysis.contentOrigin. Where the text came from: typed, pasted, or autocompleted. Has two indicators: a primary indicator for overall content source and a pasteReworkIndicator for paste editing behavior.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`charactersByOrigin.typed`	`number`	Ratio of characters typed directly (0-1, sums to 1.0 with other fields)
`charactersByOrigin.pasted`	`number`	Ratio of characters pasted (0-1, sums to 1.0 with other fields)
`charactersByOrigin.autocompleted`	`number`	Ratio of characters from autocomplete (0-1, sums to 1.0 with other fields)
`pasteEvents`	`object[]`	Individual paste events with timing, source, and rework details.
`pasteEvents[].timestamp`	`number`	Milliseconds from session start when the paste occurred.
`pasteEvents[].characterCount`	`number`	Number of characters in the pasted content.
`pasteEvents[].source`	`'internal' \| 'external' \| 'unknown'`	Whether the paste came from within the document, outside it, or is unknown.
`pasteEvents[].precedingTabAwayDuration`	`number`	Milliseconds the user was away (tab-away) immediately before this paste. (optional)
`pasteEvents[].contentMatchedDocument`	`boolean`	Whether the pasted content matched existing document text (internal restructuring).
`pasteEvents[].editsInRegion`	`number`	Number of keydowns within the paste region within 60 seconds.
`pasteEvents[].firstEditDelayMs`	`number`	Milliseconds from paste to first edit within the paste region. (optional)
`pasteEvents[].reworkRatio`	`number`	Ratio of edits to pasted chars (0.0 = untouched).
`pasteEvents[].priorClipboardOp`	`'copy' \| 'cut'`	The clipboard operation that preceded this paste (“copy” or “cut”), if known. (optional)
`pasteClassification.internalPasteRatio`	`number`	Ratio of chars from internal paste (cut-paste restructuring) to total chars (0-1)
`pasteClassification.externalPasteRatio`	`number`	Ratio of chars from external paste (outside content) to total chars (0-1)
`pasteClassification.unknownPasteRatio`	`number`	Ratio of chars from unknown-source paste to total chars (0-1)
`pasteEditSummary.pasteRetentionRate`	`number`	Fraction of pastes with zero edits in their region within 60 seconds (0-1)
`pasteEditSummary.meanModificationDelayMs`	`number`	Average delay (ms) from paste to first edit, across pastes that were edited
`pasteEditSummary.pasteToTypeRatio`	`number`	Ratio of pasted characters to typed characters
`pasteEditSummary.meanReworkRatio`	`number`	Average rework ratio across all pastes
`pasteEditSummary.pasteLengthMean`	`number`	Mean paste length in characters
`pasteEditSummary.pasteLengthStd`	`number`	Standard deviation of paste lengths
`pasteEditSummary.pasteLengthMin`	`number`	Minimum paste length in characters
`pasteEditSummary.pasteLengthMax`	`number`	Maximum paste length in characters
`contentTimeline`	`MultiTimeSeries`	Character origin ratios (typed/pasted/autocompleted) over time.

Timing & Rhythm

Access via analysis.timingAuthenticity. Whether keystroke timing looks human or mechanical. Measures dwell time, flight time distributions, periodicity, and entropy.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`dwellTimeDistribution.mean`	`number`	Mean key hold duration in milliseconds.
`dwellTimeDistribution.cv`	`number`	Variability of dwell times relative to the mean (standard deviation / mean). Higher values indicate more variation.
`dwellTimeDistribution.histogram`	`BinnedData[]`	Histogram bins for dwell time distribution.
`flightTimeDistribution.mean`	`number`	Mean inter-key interval in milliseconds.
`flightTimeDistribution.cv`	`number`	Variability of flight times relative to the mean (standard deviation / mean). Higher values indicate more variation.
`flightTimeDistribution.histogram`	`BinnedData[]`	Histogram bins for flight time distribution.
`periodicityScore`	`number`	How periodic the keystroke rhythm is (0-1, higher = more periodic/mechanical).
`entropy`	`number`	Shannon entropy of timing intervals. Higher entropy = more human-like variability.
`timingOverTime`	`MultiTimeSeries`	Timing metrics (dwell, flight) tracked across the session.

Revision Behavior

Access via analysis.revisionBehavior. How the author corrected and revised their work. Measures correction rate, edit patterns, revision depth, and insertion/deletion balance.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`correctionRate`	`number`	Ratio of correction keystrokes (Backspace/Delete) to total keystrokes (0-1).
`correctionCount`	`number`	Total number of correction keystrokes (Backspace/Delete).
`navigationCount`	`number`	Number of arrow key navigation events.
`undoRedoCount`	`number`	Number of undo/redo operations (Ctrl+Z / Ctrl+Y).
`editPatterns`	`object[]`	Categorized edit sequences with counts for each pattern type.
`editPatterns[].type`	`\| 'select-delete-type' \| 'select-paste' \| 'paste-only' \| 'select-edit' \| 'multi-step'`	The type of edit sequence observed.
`editPatterns[].count`	`number`	Number of times this edit pattern was observed.
`correctionTimeline`	`TimeSeries`	Correction rate plotted over time.
`revisionDepth.maxDepth`	`number`	Maximum number of times any single text region was revised.
`revisionDepth.avgDepth`	`number`	Average revision depth across all edited regions.
`revisionDepth.regionsRevisedMultipleTimes`	`number`	Number of distinct text regions edited more than once.
`proportionBehindFrontier`	`number`	Proportion of keydowns occurring behind the text frontier (high water mark). Higher = more revision.
`substitutedWordsCount`	`number`	Count of word-level select-then-replace events (Deane 2026).
`jumpToEditFrequency`	`number`	Ratio of non-local cursor jumps that precede an edit action (Deane 2026).
`jumpDistanceSd`	`number`	Standard deviation of cursor travel distances during non-local jumps (Deane 2026 + Conijn 2019).
`revisionAtPriorRatio`	`number`	Ratio of corrections occurring at a prior position (behind the frontier).
`revisionAtFrontierRatio`	`number`	Ratio of corrections occurring at the point of inscription (at the frontier).
`insertionDeletion.insertionCount`	`number`	Total number of insertion operations.
`insertionDeletion.totalInsertionChars`	`number`	Total characters inserted across all insertions.
`insertionDeletion.insertionLengthMean`	`number`	Mean characters per insertion.
`insertionDeletion.insertionLengthMedian`	`number`	Median characters per insertion.
`insertionDeletion.deletionCount`	`number`	Total number of deletion operations.
`insertionDeletion.totalDeletionChars`	`number`	Total characters deleted across all deletions.
`insertionDeletion.deletionLengthMean`	`number`	Mean characters per deletion.
`insertionDeletion.appendToInsertionRatio`	`number`	Ratio of appends (insertions at end) to all insertions (0-1). High = linear writing; low = revision-heavy.
`productProcessRatio`	`number`	Ratio of final text length to total characters typed. Low values indicate heavy revision; 1.0 = no deletions.
`pauseLocation.beforeSentence`	`object`	Pauses occurring before sentence boundaries.
`pauseLocation.beforeSentence.count`	`number`	Number of pauses at this location.
`pauseLocation.beforeSentence.meanDuration`	`number`	Mean pause duration in milliseconds.
`pauseLocation.beforeWord`	`object`	Pauses occurring before word boundaries.
`pauseLocation.beforeWord.count`	`number`	Number of pauses at this location.
`pauseLocation.beforeWord.meanDuration`	`number`	Mean pause duration in milliseconds.
`pauseLocation.withinWord`	`object`	Pauses occurring within words (mid-word hesitation).
`pauseLocation.withinWord.count`	`number`	Number of pauses at this location.
`pauseLocation.withinWord.meanDuration`	`number`	Mean pause duration in milliseconds.
`pauseLocation.afterWord`	`object`	Pauses occurring after word boundaries.
`pauseLocation.afterWord.count`	`number`	Number of pauses at this location.
`pauseLocation.afterWord.meanDuration`	`number`	Mean pause duration in milliseconds.
`pauseLocation.betweenWords`	`object`	Pauses occurring between words (inter-word gaps).
`pauseLocation.betweenWords.count`	`number`	Number of pauses at this location.
`pauseLocation.betweenWords.meanDuration`	`number`	Mean pause duration in milliseconds.
`proportionEventsAfterLastChar`	`number`	Proportion of keydown events where cursor position >= current text length (at the forward edge).
`proportionCharsInMultiWordInsert`	`number`	Proportion of total characters entered via multi-word paste events (2+ words).

Session Continuity

Access via analysis.sessionContinuity. Behavioral consistency throughout the session — detects abrupt rhythm shifts and tab-away patterns.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`changePoints`	`object[]`	Points where typing behavior shifted abruptly during the session.
`changePoints[].timestamp`	`number`	Milliseconds from session start when the change occurred.
`changePoints[].metric`	`string`	Which behavioral metric changed (e.g. “flightTimeMean”).
`changePoints[].valueBefore`	`number`	Metric value before the change point.
`changePoints[].valueAfter`	`number`	Metric value after the change point.
`changePoints[].magnitude`	`number`	Absolute magnitude of the change.
`tabAwayEvents`	`object[]`	Periods when the user navigated away from the tab.
`tabAwayEvents[].leftAt`	`number`	Milliseconds from session start when the user left the tab.
`tabAwayEvents[].returnedAt`	`number`	Milliseconds from session start when the user returned.
`tabAwayEvents[].duration`	`number`	Duration of the tab-away in milliseconds.
`tabAwayEvents[].activityAfterReturn`	`'typing' \| 'paste' \| 'idle'`	What the user did upon returning to the tab.
`segmentComparison`	`MultiTimeSeries`	Behavioral metrics compared across session segments over time.
`activeTime`	`ActiveTimeResult`	Active vs idle time model for this session’s continuity analysis.

Physical Plausibility

Access via analysis.physicalPlausibility. Whether the events could have been produced by a physical keyboard. Checks event timing, modifier key usage, and input method consistency.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`impossibleSequences.count`	`number`	Number of keystroke pairs faster than humanly possible.
`impossibleSequences.timeline`	`TimeSeries`	Timeline of impossible sequence occurrences.
`impossibleSequences.samples`	`object[]`	Individual impossible sequence occurrences with timestamps and durations.
`impossibleSequences.samples[].timestamp`	`number`	Milliseconds from session start when the impossible sequence occurred.
`impossibleSequences.samples[].durationMs`	`number`	Observed inter-key duration in milliseconds (below human minimum).
`syntheticEventRatio`	`number`	Ratio of events with isTrusted: false (0-1).
`mouseActivityDuringTyping.periodsWithMouseActivity`	`number`	Ratio of typing periods with recent mouse activity (0-1).
`mouseActivityDuringTyping.longestGapWithoutMouse`	`number`	Longest gap (ms) between mouse events during typing.
`zeroLatencyEventCount`	`number`	Events with zero milliseconds between them.
`unmatchedKeydownCount`	`number`	Keydown events without a corresponding keyup.
`bulkInsertCharCount`	`number`	Characters inserted in bulk rather than individual keystrokes.
`modifierEventCount`	`number`	Total shift keydown events observed.
`uncoveredCaseCount`	`number`	Uppercase characters produced without a preceding shift keydown.
`modifierLateralBalance`	`number`	Proportion of shift events that are ShiftLeft (0-1, -1 = no shifts).
`modifierCoverage`	`number`	Ratio of shift events to uppercase characters.
`modifierSustainCount`	`number`	Times shift was held across 2+ character keydowns.
`nonStandardKeys`	`object[]`	Keys not found on standard keyboard layouts, with occurrence counts.
`nonStandardKeys[].key`	`string`	The key value that is not on standard layouts.
`nonStandardKeys[].count`	`number`	Number of times this non-standard key was pressed.

Temporal Patterns

Access via analysis.temporalPatterns. Typing speed changes, fatigue, warmup, burst patterns, and pause characteristics over the session.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`sessionDurationMs`	`number`	Total session length in milliseconds.
`speedTimeline`	`TimeSeries`	Typing speed (characters per minute) plotted over time.
`fatigueRatio`	`number`	Ratio of late-session speed to early-session speed. Values < 1 indicate fatigue (slowing down).
`warmupRatio`	`number`	Ratio of initial typing speed to overall speed. Values < 1 indicate warmup (started slow).
`pauseDistribution.histogram`	`BinnedData[]`	Histogram bins for pause duration distribution.
`pauseDistribution.count`	`number`	Total number of pauses detected.
`pauseDistribution.meanDuration`	`number`	Mean pause duration in milliseconds.
`burstPattern.avgBurstLength`	`number`	Average number of keystrokes per typing burst.
`burstPattern.avgBurstSpeed`	`number`	Average typing speed (characters per minute) within bursts.
`burstPattern.burstToTotalRatio`	`number`	Ratio of keystrokes occurring within bursts to total keystrokes (0-1).
`burstPattern.burstDurationSd`	`number`	Standard deviation of burst durations (ms). High = varied composition; low = uniform transcription.
`burstPattern.burstDurationCv`	`number`	Variability of burst durations relative to the mean (standard deviation / mean).
`pause200ms`	`object`	Short pauses (>200ms). Micro-pauses reflect word-level planning.
`pause200ms.count`	`number`	Number of pauses exceeding 200ms.
`pause200ms.meanDuration`	`number`	Mean duration of 200ms+ pauses in milliseconds.
`pause200ms.sdDuration`	`number`	Standard deviation of 200ms+ pause durations.
`pause200ms.pauseRate`	`number`	Pauses per keydown (count / total keydowns).
`pause2000ms`	`object`	Long pauses (>2000ms). Cognitive pauses reflect sentence/paragraph-level planning.
`pause2000ms.count`	`number`	Number of pauses exceeding 2000ms.
`pause2000ms.meanDuration`	`number`	Mean duration of 2000ms+ pauses in milliseconds.
`pause2000ms.sdDuration`	`number`	Standard deviation of 2000ms+ pause durations.
`pause2000ms.pauseRate`	`number`	Pauses per keydown (count / total keydowns).
`shortToLongPauseRatio`	`number`	Ratio of short pauses (200ms+) to long pauses (2000ms+). Higher = more fluent typing.
`fluencySd5seg`	`number`	Standard deviation of typing speed across 5-second segments. Measures fluency variability.
`fluencySd10seg`	`number`	Standard deviation of typing speed across 10-second segments. Measures fluency variability at coarser granularity.
`burstClassification.pBurstCount`	`number`	Number of P-bursts (pause-bounded bursts between pauses >200ms).
`burstClassification.pBurstMedianDuration`	`number`	Median duration of P-bursts in milliseconds.
`burstClassification.rBurstCount`	`number`	Number of R-bursts (revision-bounded bursts between corrections).
`burstClassification.rBurstMedianDuration`	`number`	Median duration of R-bursts in milliseconds.
`burstClassification.pToRRatio`	`number`	Ratio of P-bursts to R-bursts. Reflects composition strategy.

Writing Process

Access via analysis.writingProcess. Classifies session windows into planning, drafting, and revision phases based on behavioral features.

Each category has an indicator (IndicatorOutput) with a machine-readable code and numeric params.

Metrics

Metric	Type	Description
`segments`	`WritingProcessSegment[]`	Classified segments (one per temporal window).
`transitions`	`PhaseTransition[]`	Phase transition counts and probabilities (for flow graph).
`timeInPhase`	`object`	Aggregate time-in-phase fractions.
`timeInPhase.planning`	`number`	Fraction of session time in planning (0-1).
`timeInPhase.drafting`	`number`	Fraction of session time in drafting (0-1).
`timeInPhase.revision`	`number`	Fraction of session time in revision (0-1).
`phaseChangeCount`	`number`	Number of phase changes during the session.
`windowDurationMs`	`number`	Temporal window duration used for segmentation (ms).
`phaseTimeline`	`TimeSeries`	Phase timeline: timestamps + phase index (0=planning, 1=drafting, 2=revision).

Shared Types

IndicatorOutput

Machine-readable indicator with WT-NNN code and numeric params.

Metric	Type	Description
`code`	`string`	Indicator code (e.g. “WT-100”, “WT-205”).
`params`	`Record<string, number>`	Numeric parameters specific to this indicator (e.g. pastedRatio: 0.62).

TimeSeries

Parallel arrays. All timestamps in ms from session start.

Metric	Type	Description
`timestamps`	`number[]`	Timestamps in milliseconds from session start.
`values`	`number[]`	Values corresponding to each timestamp.

MultiTimeSeries

Multiple named series sharing the same timestamp axis.

Metric	Type	Description
`timestamps`	`number[]`	Timestamps in milliseconds from session start.
`series`	`Record<string, number[]>`	Named data series, each an array of values aligned with timestamps.

BinnedData

A single bin in a histogram.

Metric	Type	Description
`binStart`	`number`	Lower bound of the bin (inclusive), in the metric’s unit.
`binEnd`	`number`	Upper bound of the bin (exclusive), in the metric’s unit.
`count`	`number`	Number of observations falling in this bin.

SessionSegment

A classified time segment within the session.

Metric	Type	Description
`startTime`	`number`	Milliseconds from session start when this segment begins.
`endTime`	`number`	Milliseconds from session start when this segment ends.
`type`	`\| 'typing' \| 'paste' \| 'pause' \| 'tabAway' \| 'autocomplete' \| 'navigating'`	The activity type during this segment.

IntegrityProof

SHA-256 hash chain over raw events for tamper detection.

Metric	Type	Description
`algorithm`	`'SHA-256'`	The hashing algorithm used.
`checkpoints`	`object[]`	—
`checkpoints[].index`	`number`	Checkpoint index (sequential from 0).
`checkpoints[].hash`	`string`	SHA-256 hash at this checkpoint.
`checkpoints[].eventCount`	`number`	Cumulative event count at this checkpoint.
`checkpoints[].timestamp`	`number`	Timestamp (ms from session start) of this checkpoint.
`finalDigest`	`string`	Final SHA-256 digest covering all events.
`eventCount`	`number`	Total number of events in the hash chain.

ActiveTimeWindow

A contiguous window of active writing time (no idle gaps).

Metric	Type	Description
`startMs`	`number`	Start of the active window in milliseconds from session start.
`endMs`	`number`	End of the active window in milliseconds from session start.

ActiveTimeResult

Active vs idle time breakdown using 3-tier gap classification (typing pause, thinking pause, idle break).

Metric	Type	Description
`activeDurationMs`	`number`	Total active writing time in milliseconds (excludes idle gaps).
`idleDurationMs`	`number`	Total idle time in milliseconds (thinking pauses + idle breaks).
`activeWindows`	`ActiveTimeWindow[]`	Contiguous windows of active writing, gaps between windows are idle time.
`thinkingPauseCount`	`number`	Number of thinking pauses detected (medium-length gaps).
`idleBreakCount`	`number`	Number of idle breaks detected (long gaps).

WindowFeatures

Per-window behavioral features used for writing process classification.

Metric	Type	Description
`cognitivePauseRate`	`number`	Cognitive pauses (>2000ms) per keydown in this window.
`burstRatio`	`number`	Fraction of keydown time spent in bursts (flight <200ms).
`correctionRate`	`number`	Corrections / keydowns in this window.
`forwardEdgeRatio`	`number`	Fraction of keydowns at or beyond the text frontier.
`cpm`	`number`	Characters per minute in this window.

WritingProcessSegment

A classified writing process segment spanning one temporal window.

Metric	Type	Description
`startTime`	`number`	ms from session start.
`endTime`	`number`	ms from session start.
`phase`	`'planning' \| 'drafting' \| 'revision'`	”planning”, “drafting”, or “revision”.
`confidence`	`number`	Classification confidence (0-1).
`features`	`WindowFeatures`	Per-window features that drove the classification.

PhaseTransition

Metric	Type	Description
`from`	`'planning' \| 'drafting' \| 'revision'`	The phase being transitioned from.
`to`	`'planning' \| 'drafting' \| 'revision'`	The phase being transitioned to.
`count`	`number`	Number of consecutive-window transitions from → to.
`probability`	`number`	Probability of this transition given `from` (0-1).