Disposable Email for LangChain & LangGraph

LangChain agents take a tools=[] list. myagentinbox exposes a tiny REST API, so the integration is three @tool decorators — one each for creating an inbox, listing messages, and reading a message.

Install

pip install langchain langchain-openai langgraph httpx

Define the tools

import httpx
from langchain_core.tools import tool

BASE = "https://myagentinbox.com"

@tool
def create_inbox() -> dict:
    """Create a disposable email inbox that expires in 24 hours.
    Returns dict with 'address' and 'expires_in'."""
    r = httpx.post(f"{BASE}/api/inboxes", timeout=10)
    r.raise_for_status()
    return r.json()["data"]

@tool
def check_inbox(address: str) -> list:
    """List messages in a disposable inbox.
    Args: address - the @myagentinbox.com email address."""
    r = httpx.get(f"{BASE}/api/inboxes/{address}/messages", timeout=10)
    r.raise_for_status()
    return r.json()["data"]

@tool
def read_message(address: str, message_id: str) -> dict:
    """Read the full body of a message.
    Args: address - inbox address. message_id - the message ID from check_inbox."""
    r = httpx.get(f"{BASE}/api/inboxes/{address}/messages/{message_id}", timeout=10)
    r.raise_for_status()
    return r.json()["data"]

email_tools = [create_inbox, check_inbox, read_message]

Use with create_react_agent (LangGraph)

from langchain_openai import ChatOpenAI
from langgraph.prebuilt import create_react_agent

llm = ChatOpenAI(model="gpt-4o")

agent = create_react_agent(
    llm,
    tools=email_tools,
    prompt=(
        "You can use disposable email. When you need an address, call create_inbox. "
        "Wait 5-10 seconds between check_inbox calls. Give up after 2 minutes."
    ),
)

result = agent.invoke({
    "messages": [{
        "role": "user",
        "content": "Sign up for example.com and read the verification email.",
    }],
})
print(result["messages"][-1].content)

Or with classic AgentExecutor

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

prompt = ChatPromptTemplate.from_messages([
    ("system", "You can read email via the provided tools."),
    ("human", "{input}"),
    ("placeholder", "{agent_scratchpad}"),
])

agent = create_tool_calling_agent(llm, email_tools, prompt)
executor = AgentExecutor(agent=agent, tools=email_tools)
executor.invoke({"input": "Create an inbox and tell me the address."})

TypeScript

import { tool } from "@langchain/core/tools";
import { z } from "zod";

const BASE = "https://myagentinbox.com";

export const createInbox = tool(
  async () => {
    const r = await fetch(`${BASE}/api/inboxes`, { method: "POST" });
    return JSON.stringify((await r.json()).data);
  },
  {
    name: "create_inbox",
    description: "Create a disposable email inbox (24h TTL).",
    schema: z.object({}),
  }
);

export const checkInbox = tool(
  async ({ address }) => {
    const r = await fetch(`${BASE}/api/inboxes/${address}/messages`);
    return JSON.stringify((await r.json()).data);
  },
  {
    name: "check_inbox",
    description: "List messages in a disposable inbox.",
    schema: z.object({ address: z.string() }),
  }
);

Limits

• Inbox creation: 3 per minute per IP
• Reads: 20 per minute per IP
• Inbox lifetime: 24 hours
• Max email size: 10 MB