Your first tracked call

Make a tracked call

Once you've installed the wrapper and configured it with your ingest token and base_url (see the quick start), every call through it is automatically tracked:

response = tracked.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "What is 2+2?"}]
)

What gets captured

The wrapper reads the response's usage field and sends usage metadata to Tokeven - among other fields, the provider and model ID, input, output, cache-read and cache-write token counts, and latency. Cost is computed server-side from those token counts using current pricing. The full field list is in Privacy architecture.

What Tokeven never receives

• •Prompt content
• •Completion content
• •Your API keys
• •File contents referenced in prompts

View in dashboard

Open Spend to see your call; it appears within seconds. If it does not, run tokeven status - it reports whether the wrapper has a token, where it is sending, and how many events are waiting in the local queue.