LangChain.js

@send0/langchain turns send0's tools into LangChain structured tools, for createAgent, model.bindTools() and LangGraph's ToolNode. It needs @langchain/core 1.x and zod 4.

InstallCopy Anchor Link

npm install @send0/langchain @langchain/core zod

ExampleCopy Anchor Link

import { ChatAnthropic } from "@langchain/anthropic";
import { send0Tools } from "@send0/langchain";
import { createAgent } from "langchain";

const agent = createAgent({
  model: new ChatAnthropic({ model: "claude-opus-5-5" }),
  tools: send0Tools({ exclude: ["send_email", "reply"] }), // reads SEND0_API_KEY
});

const result = await agent.invoke({
  messages: [{ role: "user", content: "Summarize the latest conversations in my inbox." }],
});

new Send0Toolkit(options).getTools() returns the same tools as a LangChain toolkit. A runnable version is in examples/langchain-js-agent.

OptionsCopy Anchor Link

ApprovalCopy Anchor Link

send_email and reply carry metadata.sends = true. To have a person approve them before they run, add LangChain's human-in-the-loop middleware. It pauses the agent with an interrupt, so it needs a checkpointer:

import { MemorySaver } from "@langchain/langgraph";
import { createAgent, humanInTheLoopMiddleware } from "langchain";

const agent = createAgent({
  model,
  tools: send0Tools(),
  checkpointer: new MemorySaver(),
  middleware: [humanInTheLoopMiddleware({ interruptOn: { send_email: true, reply: true } })],
});

Untrusted contentCopy Anchor Link

Every body the tools return sits inside <untrusted_email> tags with a note to treat it as data. Keep the tool set to what the task needs, use a key limited to one inbox, and put sending behind approval when the agent reads mail from strangers. See Prompt injection.