Ghost Kitchen & Virtual Brand Research
Pricing
from $0.075 / investigation bundle completed
Make sense of overlapping restaurant listings. Compare shared locations and conflicting details, then get a clear list of what to investigate next. Bring your own listings.
Customer-facing changes and compatibility notes. Current prices, source limits and recovery guidance are in the README. Preserve original files before changing a saved workflow.
Compare two sets of restaurant observations. See changed details and listings observed in only one of your supplied captures. The comparison uses each source's exact listing IDs and keeps missing identity, conflicting observations and incomplete coverage visible. It does not turn an absent listing into a confirmed closure. Ordinary single-capture review and prices are unchanged.
The complete fictional request, capture declarations, readable change brief and linked investigation questions are included. Both required comparison artifacts are acknowledged before the existing bundle charge; uncertain publication or settlement is not automatically replayed.
Updated the application build and starting guide while retaining supplied-listing analysis, market-count options, prices and limits. This does not add a listing feed, verified ownership or a new detection-accuracy claim.
Count the opportunity, not just the listings. The optional Market-Count Reality Check compares up to three grouping assumptions for a declared observation window. It keeps duplicate receipts, unassigned observations and contradictory relationships visible.
The comparison describes your supplied evidence. Physical premises and operator counts remain unknown; differences between assumptions are not statistical forecasts. The brief is included in the existing completed-analysis price, with no separate group or scenario charge.
Stronger interruption handling. A checkpoint is saved before result publication to help prevent the same interrupted run from submitting results again. A stopped worker is described as interrupted, not completed. Inspect saved results and charge information before another run; an uncertain charge may already have settled.
The guide more clearly distinguishes research clues from findings about ownership, food quality, safety or honesty. Normal prices and limits are unchanged.
Clearer subjects and evidence references in the investigation queue, a standard-header CSV starter, direct examples and pre-run help. The guide explains inconclusive results, observation dates, display limits and the completed-analysis price alongside the example.
Real-data reviews require your market, data-use permission and observation-age choice. Fictional context belongs only to the explicit example. Restored forms accept supported inactive values for retired controls; active retired settings remain unsupported. Upload time does not make an observation current.
Better treatment of conflicting evidence. Missing information becoming known is no longer treated as a contradiction. Contradictory operating observations stay visible; a newer observation does not automatically win. Copied or relabelled receipts do not inflate evidence strength. Omitted queue work is identified rather than disappearing silently.
Improved current-price checks, input preparation and recovery guidance. Results are published before a completed-analysis charge; unknown settlement is not automatically retried.
The supported product reviews supplied records. Older crawler controls and empty API input are not a way to start a new collection job. Keep the current explicit input route when updating an old saved task.
