AI Configuration
Control AI behavior for item matching, preference learning, and model selection. Changes apply globally.
Advanced Mode: Considers 'offer' vs. 'want' types, aims for direct fulfillment and complementary trades. More nuanced.
Simple Mode: Focuses on general relevance and keyword matches. Less nuanced.
Preferences Enabled (for Matching): AI matching (Advanced Mode) will consider the viewing user's motivations, location, etc., to tailor suggestions.
Auto-Inference Enabled: If enabled, users will see an option on their profile to let AI learn and (mock) update their preferences based on activity. This is experimental.
Model Selection: Chooses the default model for Genkit flows.
Note: Changes to the default model typically require an application restart to fully take effect across all backend flows.
Matching:
advanced
.
Prefs in Match:
On
.
Auto Prefs:
Off
.
Model:
Gemini 1.5 Pro (Latest)
.
AI Match Suggestion Logs
This report shows AI-generated item match suggestions. Use this log to observe how suggestions and their scores change when AI configurations (like the preferred model) are updated. The system is currently configured to use 'Gemini 1.5 Pro (Latest)'. Refresh this log after triggering new suggestions on the main site to see entries with the currently active model.
Dev Note: React Strict Mode may cause duplicate log entries in development.
Loading suggestion logs...
Live AI Preference Inference (Admin View)
Select a user to see AI-inferred trading preferences based on their structured activity data.
View Alice Trader's Profile

The AI infers preferences by analyzing structured activity data like the JSON below, generated from the user's items, current profile settings, and simulated interactions.

Live AI Preference Inference for Alice Trader:

No AI insights available for this user currently, or AI could not make suggestions.

The flow inferUserPreferencesFlow is called to perform this inference.