Choose how you want to connect
New: SDK integration now available for automatic real-time tracking!
Generate your unique Substacker API key. Keep it secret!
⚠️ Save this key now! You won't be able to see it again.
pip install substacker-sdk
npm install @substacker/sdk
Wrap your OpenAI client with our tracker. That's it!
from openai import OpenAI
from substacker import track_openai
# Wrap your OpenAI client
openai = track_openai(
OpenAI(),
api_key="YOUR_API_KEY",
team="engineering" # Your team name
)
# Use OpenAI normally - tracking happens automatically
response = openai.chat.completions.create(
model="gpt-4",
messages=[{"role": "user", "content": "Hello!"}]
)
Your application is now tracking AI costs in real-time.
Deploy your application with the SDK integration. Costs will be tracked automatically on every API call.
Visit your dashboard to see team costs, model usage, and anomaly alerts as they happen.
Configure budget thresholds and get notified when teams exceed their allocated spend.
Your API key is active. Make your first tracked API call and watch the data flow in.
Your CSV must include these columns:
model - e.g., gpt-4, gpt-3.5-turbo, claude-3-opus, gemini-proprompt_tokens - number of input tokenscompletion_tokens - number of output tokensteam - OPTIONAL - we auto-generate if missingWe can auto-generate teams from:
api_key - Maps to Team-A, Team-B, etc.user_email - Extracts from domain ([email protected] → Marketing)customer_id - Creates Customer-123, Customer-456
Example:
model,prompt_tokens,completion_tokens,team
gpt-4,100,50,Marketing
claude-3-opus,150,75,Engineering
or click to browse
CSV or JSON format • Max 50MB
Your data is processed locally and securely. We never store your raw data.
Analyzing your OpenAI usage...
Your AI costs broken down by team. Finally, you know where the money goes.
| Team | Monthly Cost | Est. CO2e | % of Total |
|---|
Upload data will appear here
You just saw your team breakdown. Now make decisions based on data.
What you can do now:
✓ Share results with your VP Eng
✓ Set team budgets based on data
✓ Plan optimization efforts
✓ Track team-level AI spending
✓ Make data-driven decisions
You'll be first in line when we launch. No spam, just results.
Cache frequent queries to eliminate duplicate API calls
Use faster, cheaper models for simple tasks
Reduce system prompt size and token usage
We can optimize your implementation and deploy the fixes in 24-48 hours.