Skip to main content

TavilySearchResults

Tavily Search is a robust search API tailored specifically for LLM Agents. It seamlessly integrates with diverse data sources to ensure a superior, relevant search experience.

This guide provides a quick overview for getting started with the Tavily search results tool. For detailed documentation of all TavilySearchResults features and configurations head to the API reference.

Overview

Integration details

ClassPackagePY supportPackage latest
TavilySearchResults@langchain/communityNPM - Version

Setup

The integration lives in the @langchain/community package, which you can install as shown below:

yarn add @langchain/community

Credentials

Set up an API key here and set it as an environment variable named TAVILY_API_KEY.

process.env.TAVILY_API_KEY = "YOUR_API_KEY";

It’s also helpful (but not needed) to set up LangSmith for best-in-class observability:

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

Instantiation

You can import and instantiate an instance of the TavilySearchResults tool like this:

import { TavilySearchResults } from "@langchain/community/tools/tavily_search";

const tool = new TavilySearchResults({
maxResults: 2,
// ...
});

Invocation

Invoke directly with args

You can invoke the tool directly like this:

await tool.invoke({
input: "what is the current weather in SF?",
});
[{"title":"San Francisco, CA Current Weather | AccuWeather","url":"https://www.accuweather.com/en/us/san-francisco/94103/current-weather/347629","content":"Current weather in San Francisco, CA. Check current conditions in San Francisco, CA with radar, hourly, and more.","score":0.9428234,"raw_content":null},{"title":"National Weather Service","url":"https://forecast.weather.gov/zipcity.php?inputstring=San+Francisco,CA","content":"NOAA National Weather Service. Current conditions at SAN FRANCISCO DOWNTOWN (SFOC1) Lat: 37.77056°NLon: 122.42694°WElev: 150.0ft.","score":0.94261247,"raw_content":null}]

Invoke with ToolCall

We can also invoke the tool with a model-generated ToolCall, in which case a ToolMessage will be returned:

// This is usually generated by a model, but we'll create a tool call directly for demo purposes.
const modelGeneratedToolCall = {
args: {
input: "what is the current weather in SF?",
},
id: "1",
name: tool.name,
type: "tool_call",
};

await tool.invoke(modelGeneratedToolCall);
ToolMessage {
"content": "[{\"title\":\"Weather in San Francisco\",\"url\":\"https://www.weatherapi.com/\",\"content\":\"{'location': {'name': 'San Francisco', 'region': 'California', 'country': 'United States of America', 'lat': 37.78, 'lon': -122.42, 'tz_id': 'America/Los_Angeles', 'localtime_epoch': 1722967498, 'localtime': '2024-08-06 11:04'}, 'current': {'last_updated_epoch': 1722967200, 'last_updated': '2024-08-06 11:00', 'temp_c': 18.4, 'temp_f': 65.2, 'is_day': 1, 'condition': {'text': 'Sunny', 'icon': '//cdn.weatherapi.com/weather/64x64/day/113.png', 'code': 1000}, 'wind_mph': 2.9, 'wind_kph': 4.7, 'wind_degree': 275, 'wind_dir': 'W', 'pressure_mb': 1015.0, 'pressure_in': 29.97, 'precip_mm': 0.0, 'precip_in': 0.0, 'humidity': 64, 'cloud': 2, 'feelslike_c': 18.5, 'feelslike_f': 65.2, 'windchill_c': 18.5, 'windchill_f': 65.2, 'heatindex_c': 18.4, 'heatindex_f': 65.2, 'dewpoint_c': 11.7, 'dewpoint_f': 53.1, 'vis_km': 10.0, 'vis_miles': 6.0, 'uv': 5.0, 'gust_mph': 4.3, 'gust_kph': 7.0}}\",\"score\":0.9983156,\"raw_content\":null},{\"title\":\"Weather in San Francisco in June 2024 - Detailed Forecast\",\"url\":\"https://www.easeweather.com/north-america/united-states/california/city-and-county-of-san-francisco/san-francisco/june\",\"content\":\"Until now, June 2024 in San Francisco is slightly cooler than the historical average by -0.6 ° C.. The forecast for June 2024 in San Francisco predicts the temperature to closely align with the historical average at 17.7 ° C. 17.7 ° C.\",\"score\":0.9905143,\"raw_content\":null}]",
"name": "tavily_search_results_json",
"additional_kwargs": {},
"response_metadata": {},
"tool_call_id": "1"
}

Chaining

We can use our tool in a chain by first binding it to a tool-calling model and then calling it:

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 { HumanMessage } from "@langchain/core/messages";
import { ChatPromptTemplate } from "@langchain/core/prompts";
import { RunnableLambda } from "@langchain/core/runnables";

const prompt = ChatPromptTemplate.fromMessages([
["system", "You are a helpful assistant."],
["placeholder", "{messages}"],
]);

const llmWithTools = llm.bindTools([tool]);

const chain = prompt.pipe(llmWithTools);

const toolChain = RunnableLambda.from(async (userInput: string, config) => {
const humanMessage = new HumanMessage(userInput);
const aiMsg = await chain.invoke(
{
messages: [new HumanMessage(userInput)],
},
config
);
const toolMsgs = await tool.batch(aiMsg.tool_calls, config);
return chain.invoke(
{
messages: [humanMessage, aiMsg, ...toolMsgs],
},
config
);
});

const toolChainResult = await toolChain.invoke(
"what is the current weather in sf?"
);
const { tool_calls, content } = toolChainResult;

console.log(
"AIMessage",
JSON.stringify(
{
tool_calls,
content,
},
null,
2
)
);
AIMessage {
"tool_calls": [],
"content": "The current weather in San Francisco is as follows:\n\n- **Condition:** Sunny\n- **Temperature:** 18.4°C (65.2°F)\n- **Wind:** 2.9 mph (4.7 kph) from the west\n- **Humidity:** 64%\n- **Visibility:** 10 km (6 miles)\n- **UV Index:** 5\n\n![Sunny](//cdn.weatherapi.com/weather/64x64/day/113.png)\n\nFor more detailed information, you can visit [WeatherAPI](https://www.weatherapi.com/)."
}

Agents

For guides on how to use LangChain tools in agents, see the LangGraph.js docs.

API reference

For detailed documentation of all TavilySearchResults features and configurations head to the API reference: https://api.js.langchain.com/classes/langchain_community_tools_tavily_search.TavilySearchResults.html


Was this page helpful?


You can also leave detailed feedback on GitHub.