About Claude API
The Claude API gives developers direct access to Claude models from their own code. It is the building block for chat features, writing tools, coding helpers and autonomous agents.
What it does
You send messages to a Claude model over HTTPS and receive a response back. You manage the conversation history yourself and decide how tools are called, so you stay in control of how your app behaves. Official client libraries exist for several programming languages, and a command-line option is also available for quick tests.
Key features
- Chat-style messages with system prompts
- Tool use so the model can call your functions
- Vision for reading images and documents
- Streaming output as the answer is written
- Prompt caching to reuse repeated context
Who it's for
Software developers and product teams who want to add Claude to a website, an internal tool or a product. It suits teams that prefer to write their own logic around the model instead of using a finished chat app.
Pricing and good to know
Usage is billed by tokens, so the cost depends on the model you pick and how much text goes in and out. API keys are created in the Claude Console. The models are also offered through Amazon Bedrock, Google Cloud and Microsoft Foundry. Check the current rates on the official site before you set a budget.
Highlights
- Send messages and receive text responses over HTTPS
- Call your own functions through tool use
- Read images and documents with vision input
- Stream responses token by token
- Reuse repeated context with prompt caching
