To use HyperAgent, you will need to connect to a langchain LLM provider. In this example, we'll be using @langchain/openai to use the gpt-4o model. For this, you'll want to setup a .env file with the OPENAI_API_KEY env var like this
OPENAI_API_KEY=sk-proj-abcd1234
Additionally, create a src directory, along with a index.ts file in it.
4
Setting up a task
The task we're running here is fairly simple. The web agent will perform a navigation to , and then perform a content extraction on that page. That extracted content will then be analyzed by gpt-4o to get the relevant content to complete task.
Insert this code into src/index.ts
import "dotenv/config";
import HyperAgent from "@hyperbrowser/agent";
import { ChatOpenAI } from "@langchain/openai";
async function runEval() {
console.log("\n===== Running Hackernews Example =====");
const llm = new ChatOpenAI({
apiKey: process.env.OPENAI_API_KEY,
model: "gpt-4o",
});
const agent = new HyperAgent({
llm: llm,
});
const result = await agent.executeTask(
"Navigate to hackernews, list me the 3 most recent articles.",
{
onStep: (step) => {
console.log(`===== STEP ${step.idx} =====`);
console.dir(step, { depth: null, colors: true });
console.log("===============\n");
},
}
);
await agent.closeAgent();
console.log("\nResult:");
console.log(result.output);
return result;
}
runEval().catch(console.error);
5
Running the task
npx ts-node src/index.ts
yarn ts-node src/index.t
And that's it! You've created and run your first web agent
