Skip to main content

OpenApiToolkit

Disclaimer ⚠️

This agent can make requests to external APIs. Use with caution, especially when granting access to users.

Be aware that this agent could theoretically send requests with provided credentials or other sensitive data to unverified or potentially malicious URLs --although it should never in theory.

Consider adding limitations to what actions can be performed via the agent, what APIs it can access, what headers can be passed, and more.

In addition, consider implementing measures to validate URLs before sending requests, and to securely handle and protect sensitive data such as credentials.

This will help you getting started with the OpenApiToolkit. For detailed documentation of all OpenApiToolkit features and configurations head to the API reference.

The OpenAPIToolkit has access to the following tools:

NameDescription
requests_getA portal to the internet. Use this when you need to get specific content from a website. Input should be a url string (i.e.Β β€œhttps://www.google.com”). The output will be the text response of the GET request.
requests_postUse this when you want to POST to a website. Input should be a json string with two keys: β€œurl” and β€œdata”. The value of β€œurl” should be a string, and the value of β€œdata” should be a dictionary of key-value pairs you want to POST to the url as a JSON body. Be careful to always use double quotes for strings in the json string. The output will be the text response of the POST request.
json_explorerCan be used to answer questions about the openapi spec for the API. Always use this tool before trying to make a request. Example inputs to this tool: β€˜What are the required query parameters for a GET request to the /bar endpoint?’ β€˜What are the required parameters in the request body for a POST request to the /foo endpoint?’ Always give this tool a specific question.

Setup​

This toolkit requires an OpenAPI spec file. The LangChain.js repository has a sample OpenAPI spec file in the examples directory. You can use this file to test the toolkit.

If you want to get automated tracing from runs of individual tools, you can also set your LangSmith API key by uncommenting below:

process.env.LANGCHAIN_TRACING_V2 = "true";
process.env.LANGCHAIN_API_KEY = "your-api-key";

Installation​

This toolkit lives in the langchain package:

yarn add langchain @langchain/core

Instantiation​

Now we can instantiate our toolkit. First, we need to define the LLM we would like to use in the toolkit.

Pick your chat model:

Install dependencies

yarn add @langchain/openai 

Add environment variables

OPENAI_API_KEY=your-api-key

Instantiate the model

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

const llm = new ChatOpenAI({
model: "gpt-4o-mini",
temperature: 0
});
import { OpenApiToolkit } from "langchain/agents/toolkits";
import * as fs from "fs";
import * as yaml from "js-yaml";
import { JsonSpec, JsonObject } from "langchain/tools";

// Load & convert the OpenAPI spec from YAML to JSON.
const yamlFile = fs.readFileSync(
"../../../../../examples/openai_openapi.yaml",
"utf8"
);
const data = yaml.load(yamlFile) as JsonObject;
if (!data) {
throw new Error("Failed to load OpenAPI spec");
}

// Define headers for the API requests.
const headers = {
"Content-Type": "application/json",
Authorization: `Bearer ${process.env.OPENAI_API_KEY}`,
};

const toolkit = new OpenApiToolkit(new JsonSpec(data), llm, headers);

Tools​

View available tools:

const tools = toolkit.getTools();

console.log(
tools.map((tool) => ({
name: tool.name,
description: tool.description,
}))
);
[
{
name: 'requests_get',
description: 'A portal to the internet. Use this when you need to get specific content from a website.\n' +
' Input should be a url string (i.e. "https://www.google.com"). The output will be the text response of the GET request.'
},
{
name: 'requests_post',
description: 'Use this when you want to POST to a website.\n' +
' Input should be a json string with two keys: "url" and "data".\n' +
' The value of "url" should be a string, and the value of "data" should be a dictionary of\n' +
' key-value pairs you want to POST to the url as a JSON body.\n' +
' Be careful to always use double quotes for strings in the json string\n' +
' The output will be the text response of the POST request.'
},
{
name: 'json_explorer',
description: '\n' +
'Can be used to answer questions about the openapi spec for the API. Always use this tool before trying to make a request. \n' +
'Example inputs to this tool: \n' +
" 'What are the required query parameters for a GET request to the /bar endpoint?'\n" +
" 'What are the required parameters in the request body for a POST request to the /foo endpoint?'\n" +
'Always give this tool a specific question.'
}
]

Use within an agent​

First, ensure you have LangGraph installed:

yarn add @langchain/langgraph
import { createReactAgent } from "@langchain/langgraph/prebuilt";

const agentExecutor = createReactAgent({ llm, tools });
const exampleQuery =
"Make a POST request to openai /chat/completions. The prompt should be 'tell me a joke.'. Ensure you use the model 'gpt-4o-mini'.";

const events = await agentExecutor.stream(
{ messages: [["user", exampleQuery]] },
{ streamMode: "values" }
);

for await (const event of events) {
const lastMsg = event.messages[event.messages.length - 1];
if (lastMsg.tool_calls?.length) {
console.dir(lastMsg.tool_calls, { depth: null });
} else if (lastMsg.content) {
console.log(lastMsg.content);
}
}
[
{
name: 'requests_post',
args: {
input: '{"url":"https://api.openai.com/v1/chat/completions","data":{"model":"gpt-4o-mini","messages":[{"role":"user","content":"tell me a joke."}]}}'
},
type: 'tool_call',
id: 'call_1HqyZrbYgKFwQRfAtsZA2uL5'
}
]
{
"id": "chatcmpl-9t36IIuRCs0WGMEy69HUqPcKvOc1w",
"object": "chat.completion",
"created": 1722906986,
"model": "gpt-4o-mini-2024-07-18",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "Why don't skeletons fight each other? \n\nThey don't have the guts!"
},
"logprobs": null,
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 12,
"completion_tokens": 15,
"total_tokens": 27
},
"system_fingerprint": "fp_48196bc67a"
}

Here's a joke for you:

**Why don't skeletons fight each other?**
They don't have the guts!

API reference​

For detailed documentation of all OpenApiToolkit features and configurations head to the API reference.


Was this page helpful?


You can also leave detailed feedback on GitHub.