Zero to agent

Why you need this

You want an agent that can actually do things — run a command, read and write files — without hand-wiring a toolset first. The ten built-in tools ship on by default, so three lines gives you a working agent. You bring only a model endpoint; Toolnexus is a library, not a runtime, so it drops into whatever control flow you already have.

The three-liner: a toolkit with the ten built-in tools, a client pointed at any OpenAI- or Anthropic-style endpoint, and one run. With no apiKey, the key is read only from the env var that belongs to the baseUrl host (openrouter.ai → OPENROUTER_API_KEY, api.openai.com → OPENAI_API_KEY, api.anthropic.com → ANTHROPIC_API_KEY, https only) and never printed. Any other host needs apiKey or apiKeyEnv — see API key by host.

JavaScript

import { createToolkit, createClient } from "toolnexus"

const tk = await createToolkit()                       // 10 built-ins, on by default
const agent = createClient({
  baseUrl: "https://openrouter.ai/api/v1", style: "openai",
  model: "openai/gpt-4o-mini", apiKey: process.env.OPENROUTER_API_KEY,
})
const res = await agent.run("List the files here, then read the largest one.", tk)
console.log(res.text)

Python

from toolnexus import create_toolkit, create_client

tk = await create_toolkit()                            # 10 built-ins, on by default
agent = create_client(
    base_url="https://openrouter.ai/api/v1", style="openai",
    model="openai/gpt-4o-mini",
)
res = await agent.run("List the files here, then read the largest one.", tk)
print(res.text)

Go

ctx := context.Background()
tk, _ := toolnexus.CreateToolkit(ctx, toolnexus.Options{})   // 10 built-ins
agent := toolnexus.CreateClient(toolnexus.ClientOptions{
    BaseURL: "https://openrouter.ai/api/v1", Style: toolnexus.StyleOpenAI,
    Model: "openai/gpt-4o-mini", APIKey: os.Getenv("OPENROUTER_API_KEY"),
})
res, _ := agent.Run(ctx, "List the files here, then read the largest one.", tk)
fmt.Println(res.Text)

Java

Toolkit tk = Toolkit.create(new Toolkit.Options());   // 10 built-ins, on by default
LlmClient agent = LlmClient.create(new LlmClient.Options()
        .baseUrl("https://openrouter.ai/api/v1")
        .style("openai")
        .model("openai/gpt-4o-mini"));
System.out.println(agent.run("List the files here, then read the largest one.", tk).text);

C#

await using var tk = await Toolkit.CreateAsync(new Toolkit.Options());   // 10 built-ins
var agent = LlmClient.Create(new LlmClient.Options {
    BaseUrl = "https://openrouter.ai/api/v1", Style = "openai",
    Model = "openai/gpt-4o-mini",
});
var res = await agent.RunAsync("List the files here, then read the largest one.", tk);
Console.WriteLine(res.Text);

Elixir

{:ok, tk} = Toolnexus.create_toolkit()                 # 10 built-ins, on by default
client = Toolnexus.Client.create(
  base_url: "https://openrouter.ai/api/v1", style: "openai",
  model: "openai/gpt-4o-mini", api_key: System.get_env("OPENROUTER_API_KEY")
)
res = Toolnexus.Client.run(client, "List the files here, then read the largest one.", tk)
IO.puts(res.text)

Clojure

(require '[toolnexus.core :as tn] '[toolnexus.client :as client])

(def tk (tn/build {:builtins true}))          ; 10 built-ins, on by default
(def agent (client/create-client
             {:base-url "https://openrouter.ai/api/v1" :style "openai"
              :model "openai/gpt-4o-mini"}))  ; api-key from OPENROUTER_API_KEY
(def res (client/run agent "List the files here, then read the largest one."
                     {:toolkit tk}))
(println (:text res))

Rust

use toolnexus::client::{ClientOptions, Style, create_client};
use toolnexus::{ToolkitOptions, create_toolkit};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let tk = create_toolkit(ToolkitOptions::default()).await?; // 10 built-ins, on by default
    let names: Vec<String> = tk.tools().into_iter().map(|t| t.name).collect();
    assert!(names.iter().any(|n| n == "bash") && names.iter().any(|n| n == "read"));
    println!("tools: {}", names.join(" "));

    if std::env::var("OPENROUTER_API_KEY").is_err() {
        println!("skip: OPENROUTER_API_KEY not set");
        return Ok(());
    }
    let agent = create_client(ClientOptions {
        base_url: "https://openrouter.ai/api/v1".into(),
        style: Style::OpenAI,
        model: "openai/gpt-4o-mini".into(), // key from OPENROUTER_API_KEY
        ..Default::default()
    });
    let res = agent.run("List the files here, then read the largest one.", &tk).await?;
    println!("{}", res.text);
    Ok(())
}

Next: MCP servers.