Manual Tracking | Paygent Documentation

Manual tracking gives you full control when you already know the exact token counts. Use manual tracking when:

✓ You’re calling AI APIs directly and have access to usage data
✓ You want maximum control over what’s reported
✓ You’re tracking non-OpenAI models or custom services
✓ You have your own token counting logic

Example: OpenAI Integration

Python

Node.js

Go

manual_openai.py

1 import os
2 import paygent_sdk
3 from paygent_sdk import UsageData
4 from openai import OpenAI
5 
6 # Initialize Paygent SDK
7 paygent = paygent_sdk.init(api_key=os.getenv("PAYGENT_API_KEY"))
8 
9 openai_client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
10 
11 # Make OpenAI API call
12 response = openai_client.chat.completions.create(
13     model="gpt-4o",
14     messages=[{"role": "user", "content": "Explain quantum physics"}]
15 )
16 
17 # Extract usage data
18 usage = response.usage
19 cached_tokens = 0
20 if hasattr(usage, 'prompt_tokens_details'):
21     cached_tokens = usage.prompt_tokens_details.cached_tokens or 0
22 
23 # Track with Paygent
24 usage_data = UsageData(
25     service_provider="OpenAI",
26     model=response.model,
27     prompt_tokens=usage.prompt_tokens,
28     completion_tokens=usage.completion_tokens,
29     cached_tokens=cached_tokens,
30     total_tokens=usage.total_tokens
31 )
32 
33 paygent.send_usage("qa-agent", "customer-456", "question-answered", usage_data)

Unified SDK Entrypoint: paygent_sdk.init(api_key) enables automatic patching for LLMs (OpenAI, Gemini, Anthropic) AND returns the paygent client object for manual tracking calls (send_usage, send_indicator, etc.).

Standalone Scripts: Paygent sends tracking events asynchronously on background threads so calling threads are never blocked. In short standalone test scripts (unlike web servers), add time.sleep(1) at the end to allow background HTTP requests to complete before the process exits.

Automate Onboarding. To simplify your workflow, call client.create_or_get_customer() before tracking usage. This ensures the customer exists in Paygent without needing manual setup in the dashboard.

Example: Gemini Integration

Python

Node.js

Go

manual_gemini.py

1 import os
2 import paygent_sdk
3 from paygent_sdk import UsageData
4 import google.generativeai as genai
5 
6 paygent = paygent_sdk.init(api_key=os.getenv("PAYGENT_API_KEY"))
7 genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
8 
9 # Make Gemini API call
10 model = genai.GenerativeModel('gemini-2.0-flash-exp')
11 response = model.generate_content('Write a poem about AI')
12 
13 # Extract usage data
14 metadata = response.usage_metadata
15 
16 # Track with Paygent
17 usage_data = UsageData(
18     service_provider="Google DeepMind",
19     model="gemini-2.0-flash-exp",
20     prompt_tokens=metadata.prompt_token_count,
21     completion_tokens=metadata.candidates_token_count,
22     total_tokens=metadata.total_token_count
23 )
24 
25 paygent.send_usage("content-gen", "customer-789", "poem-created", usage_data)