Pricing

Pricing tracks the AI costs associated with running TA training simulations. It aggregates cost data by day and model, and provides a paginated history of generation groups with token counts and dollar amounts.

What is Pricing?

Every time a TA interacts with a student persona during an office-hours simulation, Glow makes AI model calls that consume tokens and incur costs. Pricing gives administrators visibility into this spending with:

1. Daily cost chart (POST /system/pricing) — Daily aggregations showing total_cost and run_count per model, with filter options for models and agents.
2. Group history (POST /system/groups) — Paginated table of AI generation groups showing per-group token counts, costs, and associated models/agents.

This helps institutions monitor budget consumption, compare model costs, and forecast future spending for their TA training programs.

Quick Start

CLI

Calls below use $GLOW_INSTANCE_URL + $GLOW_TOKEN — see Authentication to export them once.

# Get daily cost chart data
glow system pricing
 
# Get pricing for a specific date range
glow system pricing --body '{"start_date": "2025-01-01", "end_date": "2025-01-31"}'
 
# Search group history (paginated)
glow system groups --body '{"page": 1, "page_size": 25, "sort_order": "desc"}'
 
# Export pricing data
glow system export

API

# Get daily cost aggregations
curl -X POST $GLOW_INSTANCE_URL/system/pricing \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "start_date": "2025-01-01",
    "end_date": "2025-01-31"
  }'
 
# Search group history
curl -X POST $GLOW_INSTANCE_URL/system/groups \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "page": 1,
    "page_size": 20,
    "sort_order": "desc"
  }'

Filtering

Understanding the Daily Cost Response

The POST /system/pricing response includes:

• daily — Array of PricingDailyItem buckets:
  • date_key — Date bucket (e.g., "2025-01-15")
  • model_id — Which AI model was used
  • total_cost — Dollar cost for this day + model combination
  • run_count — Number of AI runs in this bucket
• model_options — Filter dropdown options for models (value, label, count).
• agent_options — Filter dropdown options for agents.
• resources — Resource metadata keyed by ID:
  • agents — Agent display names
  • models — Model display names
• total_count — Total number of matching records.
• analytics — Filter facets for the page.

Understanding Group History

The POST /system/groups response returns paginated PricingGroupItem rows:

• group_id — UUID of the generation group
• session_id — Associated session
• group_name — Display name of the group
• first_run_at, last_run_at — Time span of the group’s runs
• run_count — Number of AI runs
• total_input_tokens — Input tokens consumed
• total_output_tokens — Output tokens generated
• total_tokens — Combined token count
• total_cost — Dollar cost as a string
• agent_ids, agent_names — Agents used
• model_ids, model_names — Models used

Model Breakdown

Exporting

# Export all pricing data as a denormalized ZIP
glow system export

The response includes content (base64-encoded), file_name, mime_type, and row_count.

Common Operations

Related

• Pricing API Reference
• Pricing CLI Reference
• Invocation Guide — individual AI model call tracking
• Group Guide — AI generation group details