LangChain Email Tool: Add Inbox and OTP Tools to Any Chain

LangChain agents can use tools to interact with external services. Adding AgentMailr tools gives your LangChain application the ability to handle email verification flows autonomously.

Define Tools Using @tool Decorator

from langchain.tools import tool
import httpx

BASE = "https://api.agentmailr.com"
HEADERS = {"X-API-Key": "ak_your_key", "Content-Type": "application/json"}

@tool
def create_inbox(name: str = "") -> dict:
    """Create a new email inbox. Returns id and address."""
    r = httpx.post(f"{BASE}/v1/inboxes", headers=HEADERS, json={"name": name})
    return r.json()

@tool
def wait_for_otp(inbox_id: str, timeout: int = 60) -> str:
    """Wait for an OTP code to arrive in an inbox. Returns the code."""
    r = httpx.get(
        f"{BASE}/v1/inboxes/{inbox_id}/otp",
        headers=HEADERS,
        params={"timeout": timeout},
        timeout=timeout + 5
    )
    data = r.json()
    return data.get("code", "No OTP received")

@tool
def send_email(inbox_id: str, to: str, subject: str, text: str) -> dict:
    """Send an email from an inbox."""
    r = httpx.post(
        f"{BASE}/v1/inboxes/{inbox_id}/send",
        headers=HEADERS,
        json={"to": to, "subject": subject, "text": text}
    )
    return r.json()

Create an Agent with These Tools

from langchain.agents import create_tool_calling_agent, AgentExecutor
from langchain_anthropic import ChatAnthropic
from langchain_core.prompts import ChatPromptTemplate

llm = ChatAnthropic(model="claude-sonnet-4-6")
tools = [create_inbox, wait_for_otp, send_email]

prompt = ChatPromptTemplate.from_messages([
    ("system", "You are an agent that can manage email inboxes and handle verification flows."),
    ("human", "{input}"),
    ("placeholder", "{agent_scratchpad}"),
])

agent = create_tool_calling_agent(llm, tools, prompt)
executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

result = executor.invoke({
    "input": "Create an inbox, then send a test email to demo@example.com from it."
})

Use in Chains

These tools work in any LangChain chain or agent: ReAct agents, OpenAI Functions agents, and LCEL chains with tool calling. The same tool definitions work with both Python and JavaScript LangChain.