Analysis Job
An Analysis Job is the traceable processing record created each time you submit a video or session for analysis. This means it tracks status (queued, running, complete, failed), the inputs used. The outputs produced.
Every analysis in SwingVantage runs as an identifiable job with a lifecycle. Jobs are stored locally first (so they survive network interruptions) and synced to the cloud when connected. Admins can inspect job logs to debug failures or audit AI usage. For athletes, the job system means you can always find the inputs and outputs of any past analysis. This means what video was submitted, what the engine returned, and what confidence it had. Instead of losing that context over time.
If an analysis takes unusually long, check the job log status rather than submitting the same video again. Duplicate submissions create duplicate jobs and consume your analysis quota without providing additional information.
Use the job log to review AI token usage per analysis session. This means if costs are running high, look for sessions where AI enhancement ran on low-quality video (low confidence scores). This produces expensive but low-value AI output that can be avoided with a quality gate check first.
Example
Analysis Job #4782: submitted 6-iron video, 2026-06-28 14:32 — Status: complete — Primary fault: over-the-top (Confidence: 74/100) — AI layer: used (0.4K tokens).
How it shows up on video
If your analysis job shows a "failed" status in the job log, check the video quality gate score for that upload first. This means most job failures are traceable to a video that barely passed the gate and then degraded at critical frames during processing. This produces a result the engine could not complete.
Common mistakes
- Submitting the same video twice when a job is running slowly — duplicate submissions create duplicate jobs and consume processing quota without producing additional information.
- Not checking the job log when results seem wrong. This means the job log shows exactly which video frames triggered which findings, letting you verify or dispute a result with evidence.
- Assuming a completed job always produced reliable results. This means job completion means the pipeline ran to end. It does not mean the video quality was sufficient for high-confidence output.
- Deleting old job records before reviewing them. This means job history is the audit trail for your analysis provenance. Retaining it allows you to re-examine earlier findings as your understanding improves.
In SwingVantage Motion Lab
Each analysis job populates Motion Lab with its outputs — the 3D avatar, skeleton overlay, metrics, and fault annotations all come from a single job run. If Motion Lab shows unexpected results, the job log identifies exactly which video frames triggered each finding, providing a traceable evidence path.
Related terms
- AI Swing ReportThe AI Swing Report is the comprehensive analysis document SwingVantage generates after processing your video, combining the deterministic engine's findings with optional AI-enhanced commentary into a single, prioritized report.
- Heuristic EngineThe heuristic engine is the deterministic, rules-based part of SwingVantage that runs first on every analysis — no external AI call required.
- Confidence ScoreA confidence score is a 0–100 calibration of how much to trust a finding, scaled by sample size, shot-to-shot consistency, and how complete your inputs were.
- Video Quality GateThe Video Quality Gate is the automatic pre-analysis check that evaluates a submitted clip's usability and blocks or warns before processing video that is too low-quality to produce reliable analysis.
Put this into your swing
SwingVantage can spot this in your own swing — free to start.