PondralWhat we shipped.
Release notes, engine coverage updates, methodology changes. Updated weekly.
Audit finalize transaction repair
Patched the JSON encoding on the finalize_audit_metrics RPC. 18 prior audit runs had failed to write their aeo_score. New runs now persist the run-level score and the per-engine breakdown cleanly.
Grok 4 support
Added support for xAI's Grok 4. Backfilled all active workspaces with a 90-day historical baseline.
Linear & Asana integration
Convert any gap into a tracked task with one click. Available on Growth and Agency.
Citation link detection improvements
Fixed false-negatives where models reference a brand without a hyperlink but with a clearly attributed quote.
Published methodology v2 with five-factor rubric
Methodology v2 ships with the five-factor rubric (Presence, Prominence, Context, Citation Link, Competitive Presence). Scores are the mean of one graded response per (query, engine) pair; repeated sampling across independent trials is available on request. Inter-rater agreement statistics, multi-run averaging by default, and confidence intervals are tracked on the methodology v3 roadmap.