Skip to main content

​
Overview

The get_spend_history action retrieves and analyzes your GPU rental spending history from the Hyperbolic platform. It provides detailed information about past rentals, costs, and usage statistics.

​
Function Signature

def get_spend_history() -> str

​
Input Schema

This action takes no input parameters.

​
Response

Returns a formatted analysis of your GPU rental spending.
​
analysis
string
A formatted string containing:
  • List of all instances rented
  • Duration of each rental in seconds
  • Cost per rental in USD
  • Total spending per GPU type
  • Overall total spending

​
API Endpoint

GET https://api.hyperbolic.xyz/v1/marketplace/instances/history

​
Example Response

=== GPU Rental Spending Analysis ===

Instance Rentals:
- respectful-rose-pelican:
  GPU: NVIDIA H100 (Count: 2)
  Duration: 3600 seconds
  Cost: $5.00
- happy-blue-dolphin:
  GPU: NVIDIA A100 (Count: 1)
  Duration: 7200 seconds
  Cost: $3.50

GPU Type Statistics:

NVIDIA H100:
  Total Rentals: 2
  Total Time: 3600 seconds
  Total Cost: $5.00

NVIDIA A100:
  Total Rentals: 1
  Total Time: 7200 seconds
  Total Cost: $3.50

Total Spending: $8.50

​
Example Usage

from hyperbolic_agentkit_core.actions import get_spend_history

# Get spending history
history = get_spend_history()
print(history)

​
Response Details

The spending analysis includes:

​
Instance Rentals

For each rented instance:
  • Instance name: Unique identifier for the rental
  • GPU model: Type of GPU used (e.g., NVIDIA H100, A100)
  • GPU count: Number of GPUs in the instance
  • Duration: Rental duration in seconds
  • Cost: Total cost in USD

​
GPU Type Statistics

Aggregated by GPU model:
  • Total Rentals: Number of GPUs of this type rented
  • Total Time: Combined rental duration in seconds
  • Total Cost: Combined spending in USD

​
Summary

  • Total Spending: Overall spending across all rentals in USD

​
Cost Calculation

Costs are calculated using the formula: 
cost = (duration_hours × price_per_hour_in_cents) / 100
Where:
  • duration_hours = duration_seconds / 3600
  • price_per_hour_in_cents is from the instance pricing data
  • Result is converted to dollars by dividing by 100

​
Important Notes

  • Requires HYPERBOLIC_API_KEY environment variable
  • Returns “No rental history found.” if you haven’t rented any instances
  • All costs are in USD
  • Durations are in seconds for precision
  • Only includes terminated instances (active rentals appear in get_gpu_status)
  • Statistics are grouped by GPU model for easy analysis

​
Error Handling

from hyperbolic_agentkit_core.actions import get_spend_history

try:
    history = get_spend_history()
    print(history)
except Exception as e:
    print(f"Error retrieving spend history: {e}")
If the API request fails, returns: 
Error retrieving spend history: [error details]

​
Use Cases

​
Budget Tracking

Monitor your GPU rental spending to stay within budget: 
history = get_spend_history()
if "Total Spending: $" in history:
    # Extract and check total
    pass

​
Usage Analysis

Analyze which GPU types you use most: 
# The response includes per-GPU statistics
history = get_spend_history()
print(history)

​
Cost Optimization

Compare costs across different GPU types to optimize spending: 
# Review GPU Type Statistics section
# Compare total cost vs. total time for different models

Build docs developers (and LLMs) love

Get started for freeTalk to us