Usage Analytics

Fresh

Source: docs.hindsight.vectorize.io/usage-analytics

Dashboard

Navigate to Organization > Usage to access real-time monitoring.

Time Periods

Period	Resolution	Use Case
24 Hours	Hourly	Recent activity, debugging
7 Days	Daily	Weekly patterns
30 Days	Daily	Monthly trends

Dashboard auto-refreshes every 30 seconds.

Summary Cards

Metric	Color	Description
Total Tokens	—	Combined usage
Retain Tokens	Blue	Memory storage
Recall Tokens	Green	Memory retrieval
Reflect Tokens	Purple	AI insights
Retrieve Model	Teal	Mental model GET
Refresh Model	Orange	Mental model create/refresh
File Conversion (Iris)	Red	Enhanced extraction

Optimization Strategies

Retain: Batch multiple memories, avoid redundancy, use efficient text formats

Recall: Use targeted queries, apply type filters, cache frequent results

Reflect: Clear questions, limit context, reuse insights

Mental Models: Prefer retrieval (cheap) over refresh (runs Reflect internally)

File Conversion: Use Standard (free) for text documents; reserve Enhanced (Iris) for images and scanned PDFs

Reference

• 1 million tokens is approximately 750,000 words
• Average word consumes 1-2 tokens
• Real-time accuracy within minutes