Count Tokens

The /v1/messages/count_tokens endpoint delegates directly to the upstream Anthropic count_tokens endpoint with $0 inference spend. It calculates the exact input token count an Anthropic Messages payload will consume when sent to Claude models, without generating tokens or reserving credits.

nRouter validates key permissions and proxies the payload (system, messages, multimodal data, and tools) to the authoritative upstream tokenizer. The integer count is returned with no inference charges.

POST https://api.nrouter.ai/v1/messages/count_tokens

Request Parameters

Examples

import { nRouter } from "@nrouter_ai/sdk";
const client = new nRouter({ apiKey: process.env.NROUTER_API_KEY });
const count = await client.messages.countTokens({
  model: "claude-sonnet-4-5-20250929",
  messages: [{ role: "user", content: "Count tokens for this sentence." }],
});
console.log("Tokens:", count.input_tokens);

Headers & Errors

Inbound: Authorization: Bearer sk-nrouter-... and Content-Type: application/json. Outbound: x-nr-request-id, x-nr-latency-ms, x-nr-model, x-nr-guardrails. See Chat Completions.

curl -X POST "https://example.com/v1/messages/count_tokens" \  -H "Content-Type: application/json" \  -d '{    "model": "string",    "messages": [      {}    ]  }'

{  "input_tokens": 9}