openapi: 3.2.0
info:
  title: LiteLLM [beta] Anthropic Messages Token Counting API
  description: "Enterprise Edition \n\nProxy Server to call 100+ LLMs in the OpenAI format. [**Customize Swagger Docs**](https://docs.litellm.ai/docs/proxy/enterprise#swagger-docs---custom-routes--branding)\n\n\U0001F449 [```LiteLLM Admin Panel on /ui```](/ui). Create, Edit Keys with SSO. Having issues? Try [```Fallback Login```](/fallback/login)\n\n\U0001F4B8 [```LiteLLM Model Cost Map```](https://models.litellm.ai/).\n\n\U0001F50E [```LiteLLM Model Hub```](/ui/model_hub_table). See available models on the proxy. [**Docs**](https://docs.litellm.ai/docs/proxy/ai_hub)"
  version: 1.95.0
  x-operator: institution
  x-provenance:
    method: probed
    source: https://llmproxy.uva.nl/openapi.json
    retrieved: '2026-08-19'
    note: Document is generated by the LiteLLM proxy software the University of Amsterdam self-hosts; the deployment, the key issuance and the host (llmproxy.uva.nl, UvA Azure) are the institution's. servers[] added by API Evangelist because the served document omits it; nothing else altered.
servers:
- url: https://llmproxy.uva.nl
  description: University of Amsterdam / Amsterdam University of Applied Sciences shared AI gateway
tags:
- name: '[beta] Anthropic Messages Token Counting'
paths:
  /v1/messages/count_tokens:
    post:
      tags:
      - '[beta] Anthropic Messages Token Counting'
      summary: Count Tokens
      description: "Count tokens for Anthropic Messages API format.\n\nThis endpoint follows the Anthropic Messages API token counting specification.\nIt accepts the same parameters as the /v1/messages endpoint but returns\ntoken counts instead of generating a response.\n\nExample usage:\n```\ncurl -X POST \"http://localhost:4000/v1/messages/count_tokens?beta=true\"       -H \"Content-Type: application/json\"       -H \"Authorization: Bearer your-key\"       -d '{\n    \"model\": \"claude-3-sonnet-20240229\",\n    \"messages\": [{\"role\": \"user\", \"content\": \"Hello Claude!\"}]\n  }'\n```\n\nReturns: {\"input_tokens\": <number>}"
      operationId: count_tokens_v1_messages_count_tokens_post
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema: {}
      security:
      - APIKeyHeader: []
components:
  securitySchemes:
    APIKeyHeader:
      type: apiKey
      description: Bearer token
      in: header
      name: x-litellm-api-key