Skip to Content
Model API ReferenceAnthropic MessagesToken counting

Count input tokens

POST https://api.fenno.ai/v1/messages/count_tokens
x-api-key: <FENNO_API_KEY>
anthropic-version: 2023-06-01
Content-Type: application/json

This endpoint estimates input tokens without generating an answer. Availability can depend on the model and request blocks; it does not guarantee that every Messages model can count every input type.

Request and response

Send model and the messages you intend to use. Include relevant system and tools definitions where supported so the estimate reflects the actual request. Do not add generation’s required max_tokens to this counting example.

import os
from anthropic import Anthropic
 
client = Anthropic(
    base_url="https://api.fenno.ai",
    api_key=os.environ["FENNO_API_KEY"],
)
count = client.messages.count_tokens(
    model=os.environ["FENNO_MODEL"],
    messages=[{"role": "user", "content": "What is an API gateway?"}],
)
print(count.input_tokens)

The response includes input_tokens. It is not an output-token prediction or a currency bill. Final usage can differ because generation, tool rounds and model-side processing change the total workload.

Use estimates safely

  • Check input size before starting an expensive workflow, while reserving room for the intended output.
  • If a model or block type is unsupported, surface that limitation rather than returning a fabricated zero.
  • Local tokenizers are approximate unless explicitly matched to the selected model and request format.
  • Keep pricing calculations separate and reconcile actual usage with available billing records.

See Messages and cost management.

MessagesOverview