Messages API, tools, vision, streaming SSE, prompt caching

Claude API by Anthropic - Messages, Tools, Vision, Streaming

Use the Messages API to send structured message lists with text and image blocks, stream responses over SSE, and connect tools for reliable workflows.

Choose a Claude Model

Pick the model that matches your latency, cost, and reasoning needs.

Claude Haiku 4.5

Fast response and low cost for high-frequency calls and lightweight tasks.

View model details

Claude Sonnet 4.5

Balanced performance and cost for general-purpose chat, analysis, and content workflows.

View model details

Claude Opus 4.1

Stable version for complex reasoning tasks and production-grade use cases.

View model details

Claude Sonnet 4.0

Classic version with reliable performance for everyday workloads.

View model details

Claude Opus 4.5

Strongest capabilities for deep analysis, planning, and multi-step workflows.

View model details

Online Demo (Qwen)

This demo showcases the message workflow only and does not use Claude models.

Note: The demo below is for interaction flow demonstration. For production, select a Claude model and call the Claude API.

Choose Claude model

Loading demo...

What is Claude API?

Claude API is Anthropic's developer platform for building with Claude models through the Messages API, supporting structured prompts, multimodal input, and tool use.

Messages API

Send a messages array to /v1/messages with a model and max_tokens. The API is stateless, so include the full conversation history for multi-turn apps.

Vision and Multimodal Input

Include text and image blocks together. Images support base64 or URL sources and image/jpeg, image/png, image/gif, and image/webp types.

Streaming Responses (SSE)

Set stream: true to receive server-sent events and render responses incrementally.

Tool Use with JSON Schema

Define tools in the tools parameter with name, description, and input_schema (JSON Schema) so Claude can call your functions.

Why Choose Claude API?

Modern API primitives designed for reliable production apps and agent workflows.

Stateless, Replayable Conversations

Messages are stateless. Send the full history each turn to keep multi-step workflows deterministic and easy to debug.

Prompt Caching for Reuse

Add cache_control to reuse prompt prefixes. Default cache lifetime is 5 minutes with an optional 1 hour cache for repeated, long prompts.

Streaming SSE for Real-Time UX

Stream output over SSE for chat and assistant experiences with incremental output.

Tool-Ready Automation

Define tools with JSON Schema input and let Claude call them when needed.

Claude API Use Cases

From content generation to agent automation, Claude API supports end-to-end workflows.

Customer support automation

Summarize tickets, draft replies, and keep brand tone consistent across channels.

Placeholder image

Document and image Q&A

Analyze documents or images by combining text with vision input in one request.

Placeholder image

Code review and refactoring

Explain code, flag risks, and propose safer, cleaner implementations with rationale.

Placeholder image

Structured data extraction

Transform unstructured text into JSON for downstream analytics and automation.

Placeholder image

Agent workflows

Combine tools, system prompts, and multi-turn messages to automate complex tasks.

Placeholder image

Localization and rewriting

Translate and adapt product content while keeping terminology and tone consistent.

Placeholder image

Quick Start Guide

From API key to your first Messages request in minutes.

  • Create an API Key

    Create an API key in the Console and store it securely.

  • Call /v1/messages

    Send a POST request with model, max_tokens, and a messages array. Include the full history for multi-turn conversations.

  • Add streaming, tools, and images

    Enable stream: true for SSE, define tools with JSON Schema, and include image blocks (base64 or URL) when needed.

curl --request POST \
  --url https://api.evolink.ai/v1/messages \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, world" }
    ]
  }'

Frequently Asked Questions

Ready to Start with Claude API?

Pick the right model, get your API key, and ship your first Messages request today.

Choose a model