Utilities

Helper functions for model information and token estimation.

getModelInfo()

Retrieves information about a specific model's capabilities and pricing.

Signature:

function getModelInfo(modelId: string): ModelInfo | undefined

Parameters:

Returns: ModelInfo | undefined — Model information if found, undefined otherwise.

Example:

import { getModelInfo } from '@motioneffector/llm'

const info = getModelInfo('anthropic/claude-sonnet-4')

if (info) {
  console.log(`Context: ${info.contextLength} tokens`)
  console.log(`Prompt cost: ${info.pricing.prompt}/M tokens`)
  console.log(`Completion cost: ${info.pricing.completion}/M tokens`)
}

Available Models:

Notes:

• Returns undefined for unknown models
• Pricing is in USD per million tokens
• Data may not reflect current API pricing; check provider for latest rates

estimateTokens()

Estimates the number of tokens in a text string.

Signature:

function estimateTokens(text: string): number

Parameters:

Returns: number — Estimated token count.

Example:

import { estimateTokens } from '@motioneffector/llm'

const text = 'Hello, world! How are you doing today?'
const tokens = estimateTokens(text)

console.log(`Estimated tokens: ${tokens}`)
// Estimated tokens: 10

Notes:

• Uses a simple heuristic: 1 token ≈ 4 characters
• This is an approximation; actual tokenization varies by model
• Empty strings return 0
• Useful for rough estimates before sending requests

Example with messages:

import { estimateTokens } from '@motioneffector/llm'

const messages = [
  { role: 'system', content: 'You are a helpful assistant.' },
  { role: 'user', content: 'Explain quantum computing.' }
]

let totalTokens = 0
for (const message of messages) {
  totalTokens += estimateTokens(message.content)
  totalTokens += 3  // Overhead per message
}

console.log(`Estimated prompt tokens: ${totalTokens}`)

Types

ModelInfo

interface ModelInfo {
  contextLength: number
  pricing: {
    prompt: number
    completion: number
  }
}