Anthropic Agent SDK

Build tool-use agents with Claude using the Anthropic SDK. Combine with AICredits for billing, monitoring, and multi-model support.

The native Anthropic SDK gives you access to Claude-specific features like extended thinking and fine-grained cache control. For most agent use cases, the OpenAI-compatible AICredits endpoint is simpler and lets you switch models freely.

Overview

You have two ways to use Claude with AICredits:

Approach	When to use	Base URL
AICredits OpenAI endpoint	Most agent use cases — tool use, streaming, RAG	`https://api.aicredits.in/v1`
Native Anthropic SDK via AICredits	Extended thinking, explicit cache_control	`https://api.aicredits.in` (no `/v1`)

When to Use the Native SDK

Use the native Anthropic SDK only when you need Claude-exclusive features not available through the OpenAI-compatible layer:

Extended thinking / deep reasoning (budget_tokens)
Explicit cache_control markers for prompt caching
Vision with fine-grained image block control
Computer use (beta)

For everything else — tool use, structured output, streaming, RAG — use the AICredits OpenAI-compatible endpoint with Claude model IDs.

Tool Use Loop (Native SDK)

import json
from anthropic import Anthropic

# base_url without /v1 — the Anthropic SDK appends /v1/messages internally
client = Anthropic(
    base_url="https://api.aicredits.in",
    api_key="sk-your-aicredits-key-here",
)

tools = [
    {
        "name": "search_docs",
        "description": "Search the documentation for information",
        "input_schema": {
            "type": "object",
            "properties": {
                "query": {"type": "string", "description": "Search query"},
            },
            "required": ["query"],
        },
    }
]

def search_docs(query: str) -> str:
    return f"Results for '{query}': AICredits supports 300+ models..."

messages = [{"role": "user", "content": "What models does AICredits support?"}]

while True:
    response = client.messages.create(
        model="anthropic/claude-sonnet-4.5",
        max_tokens=1024,
        tools=tools,
        messages=messages,
    )

    messages.append({"role": "assistant", "content": response.content})

    if response.stop_reason == "end_turn":
        for block in response.content:
            if block.type == "text":
                print(block.text)
        break

    if response.stop_reason == "tool_use":
        tool_results = []
        for block in response.content:
            if block.type == "tool_use":
                result = search_docs(**block.input)
                tool_results.append({
                    "type": "tool_result",
                    "tool_use_id": block.id,
                    "content": result,
                })

        messages.append({"role": "user", "content": tool_results})

Claude via AICredits (OpenAI SDK)

For the majority of Claude agent use cases, the OpenAI SDK + AICredits endpoint is simpler and gives you unified billing across all providers:

import json
from openai import OpenAI

client = OpenAI(
    base_url="https://api.aicredits.in/v1",
    api_key="sk-your-key-here",
)

tools = [{
    "type": "function",
    "function": {
        "name": "search_docs",
        "description": "Search the documentation for information",
        "parameters": {
            "type": "object",
            "properties": {
                "query": {"type": "string"},
            },
            "required": ["query"],
        },
    },
}]

messages = [{"role": "user", "content": "What models does AICredits support?"}]

while True:
    response = client.chat.completions.create(
        model="anthropic/claude-sonnet-4.5",
        messages=messages,
        tools=tools,
        tool_choice="auto",
    )

    message = response.choices[0].message
    messages.append(message)

    if response.choices[0].finish_reason == "stop":
        print(message.content)
        break

    if message.tool_calls:
        for tool_call in message.tool_calls:
            args = json.loads(tool_call.function.arguments)
            result = "AICredits supports 300+ models from OpenAI, Anthropic, Google..."
            messages.append({
                "role": "tool",
                "tool_call_id": tool_call.id,
                "content": result,
            })

Hybrid Approach

Use the native SDK for thinking features and AICredits for everything else:

from anthropic import Anthropic
from openai import OpenAI

# Anthropic SDK: base_url WITHOUT /v1 — SDK appends /v1/messages internally
anthropic_client = Anthropic(
    base_url="https://api.aicredits.in",
    api_key="sk-your-aicredits-key-here",
)

# OpenAI SDK: base_url WITH /v1 — SDK appends /chat/completions
aicredits_client = OpenAI(
    base_url="https://api.aicredits.in/v1",
    api_key="sk-your-aicredits-key-here",
)

def plan_with_thinking(task: str) -> str:
    """Use extended thinking to plan the approach."""
    response = anthropic_client.messages.create(
        model="anthropic/claude-sonnet-4.5",
        max_tokens=8000,
        thinking={"type": "enabled", "budget_tokens": 5000},
        messages=[{"role": "user", "content": f"Plan how to: {task}"}],
    )
    return next(b.text for b in response.content if b.type == "text")

def execute_plan(plan: str, task: str) -> str:
    """Execute the plan through AICredits (billed to your wallet)."""
    response = aicredits_client.chat.completions.create(
        model="anthropic/claude-sonnet-4.5",
        messages=[
            {"role": "system", "content": f"Execute this plan: {plan}"},
            {"role": "user", "content": task},
        ],
    )
    return response.choices[0].message.content

Base URL reminder: The Anthropic SDK uses https://api.aicredits.in (no /v1) — the SDK appends /v1/messages internally. The OpenAI SDK uses https://api.aicredits.in/v1 (with /v1) — it appends /chat/completions. Both route all billing through your AICredits INR wallet.

Overview

When to Use the Native SDK

Tool Use Loop (Native SDK)

Claude via AICredits (OpenAI SDK)

Hybrid Approach

On this page