Skip to content
Eko Docs

Quickstart

Here are two quick ways to get started:

  1. Using a browser extension: Suitable for those who just want to try it out or are not professionals.
  2. Running a Node.js script: Suitable for professionals who want to review or modify the code details.

Using a browser extension

Let’s run an Eko workflow together in a browser extension to automate the task that Open Twitter, search for "Fellou AI" and follow.

Load extension

  • Download the precompiled extension (or you can also build it yourself). Unzip the ZIP file to a suitable location, and you should see a dist folder.
  • Open the Chrome browser and navigate to chrome://extensions/.
  • Turn on Developer mode (toggle switch in the top right corner).
  • Click Load unpacked button (the blue text in the top-left corner) and select the dist folder in the first step.

Configure LLM model API Key

  • If it’s inconvenient to obtain an API key from the OpenAI or Claude platform, consider using proxy sites or services (such as OpenRouter), and then replace the Base URL and API key with the corresponding values.

Let’s run it!

Open the side-bar of the extension, input your prompt, and click the Run button:

Running a Node.js script

First we need to create a new project:

Terminal window
mkdir try-eko
cd try-eko
npm init

And install dependencies:

Terminal window
npm add @eko-ai/eko @eko-ai/eko-nodejs ts-node

Then write a script named index.ts:

index.ts
import { Eko, Agent, LLMs } from "@eko-ai/eko";
import { BrowserAgent } from "@eko-ai/eko-nodejs";


async function run() {
  let llms: LLMs = {
    default: {
      provider: "anthropic",
      model: "claude-3-5-sonnet-20241022",
      apiKey: "sk-xxx", // replace it with your API KEY
      config: {
        baseURL: "https://api.anthropic.com/v1",
      },
    },
  };
  let agents: Agent[] = [new BrowserAgent()];
  let eko = new Eko({ llms, agents });
  let result = await eko.run("Search for the latest news about Musk");
  console.log("result: ", result.result);
}


run().catch(e => {
  console.log(e)
});

Remember to set the environment variables (one of OpenAI/Claude):

Terminal window
export OPENAI_BASE_URL=your_value
export OPENAI_API_KEY=your_value
export ANTHROPIC_BASE_URL=your_value
export ANTHROPIC_API_KEY=your_value

Finally run it:

Terminal window
ts-node index.ts

Next Steps

Now that you have run the first workflow, you can: