THE CLASSIFIER

A$1 / 1,000,000,000 INPUT TOKENS

For all the things you're using an LLM for
that aren't language generation.

Documentation · Models · Get API Key · Pricing

WHAT IT DOES

Route an agent.

Classify a ticket.

Score an LLM response.

Rank a lead.

Detect an objection.

Choose between fixed alternatives.

CALL

result = tsnli.classify(
    text="Please cancel my account at the end of the month.",
    choices=[
        "billing",
        "technical support",
        "cancellation",
        "fraud"
    ]
)

Response:

{
  "cancellation": 0.982,
  "billing": 0.011,
  "technical support": 0.004,
  "fraud": 0.003
}

The HTTP call is POST https://api.theclassifier.ai/v1/classify. The small client is in client/tsnli.py.

PRICING

The monthly price is the credits included with the plan. Credits are spent at A$1 per 1,000,000,000 input tokens.

There is no way to add credits. When they run out, the next plan up is required.

We charge for access to the machine. Tokens are basically free.

Why is it this cheap?
Because classification is not generation.

It classifies text.

You send text and some possible answers. It returns probabilities.

It is fast. It is cheap. Use it if it works better than whatever you're currently using.