LangChain.js

A LangChain.js agent runs on Rebuno unchanged apart from two seams. The chat model takes rebunoFetch, so every model call is recorded as an llm_call step, and tools run through defineTool, so every tool call is recorded as a tool_call step. On a re-dispatch the agent runs from the top and recorded steps replay instead of calling the model or the tool again.

Install

npm install rebuno langchain @langchain/core @langchain/openai zod

Model calls

ChatOpenAI passes configuration through to the OpenAI client, which accepts a custom fetch. Pass rebunoFetch:

import { ChatOpenAI } from "@langchain/openai";
import { rebunoFetch } from "rebuno";

const model = new ChatOpenAI({ model: "gpt-5.5", configuration: { fetch: rebunoFetch } });

See LLM calls.

Tools

Declare the Rebuno side with defineTool, and call it from a LangChain tool:

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

const sendEmail = defineTool({
  name: "send_email",
  idempotency: "at_most_once",
  execute: async ({ body }: { body: string }) => mail.send("ops@acme.com", body),
});

const sendEmailTool = tool((args) => sendEmail(args), {
  name: "send_email",
  description: "Email the support summary.",
  schema: z.object({ body: z.string() }),
});

Mark anything with a side effect, such as sending an email or creating a ticket, at_most_once. See idempotency.

Stopping on approval

When a tool is held for approval, it throws. The agent loop passes a tool error back to the model and keeps looping, and every model call after that is refused. Abort the loop on the first tool error instead, by catching it in the tool and aborting a signal passed to invoke. The agent below does this.

The agent

import { createAgent } from "langchain";
import { Agent } from "rebuno";

async function process(input: { query: string }) {
  const model = new ChatOpenAI({ model: "gpt-5.5", configuration: { fetch: rebunoFetch } });
  const halt = new AbortController();
  const tools = [
    tool(
      async (args) => {
        try { return await sendEmail(args); }
        catch (error) { halt.abort(error); throw error; }
      },
      { name: "send_email", description: "Email the support summary.", schema: z.object({ body: z.string() }) },
    ),
  ];

  const graph = createAgent({ model, tools });
  const result = await graph.invoke(
    { messages: [{ role: "user", content: input.query }] },
    { signal: halt.signal },
  );
  return { answer: result.messages.at(-1)?.content };
}

const agent = new Agent("support");
await agent.serve({ port: 5000 }, process);

What Rebuno adds

Policy. Every model and tool call is checked against policy before it runs. A denied tool returns the rule’s reason to the model as the tool result, so the agent can take a different path.
Approvals. A tool that requires approval parks the execution. Once it’s approved, the agent is dispatched again, the earlier steps replay, and the approved call runs.
Recovery. If the worker dies partway through, the next dispatch replays every completed step and continues from the first one that didn’t finish. Model calls that already ran are not paid for twice.

Full example

examples/frameworks/typescript/langchain_agent.ts is a support agent that investigates a customer issue, creates a ticket, and emails a summary after approval.

Get started

Writing agents

Frameworks

Reference

Install

Model calls

Tools

Stopping on approval

The agent

What Rebuno adds

Full example

​Install

​Model calls

​Tools

​Stopping on approval

​The agent

​What Rebuno adds

​Full example

Install

Model calls

Tools

Stopping on approval

The agent

What Rebuno adds

Full example