Install

Your API key

Demo key

sk-abto-live-P4c…Create key

Shown once at creation.

First user request received.See it in Overview

1. Route your LLM calls

secret key

Send LLM traffic through the ABTO gateway so cost, latency, and results are captured at the source. Use your sk-abto-… key (server-side).

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.ABTO_API_KEY, // your ABTO secret key
  baseURL: "https://api.abto.app/v1", // ABTO gateway
});

const completion = await client.chat.completions.create({
  model: "gpt-4o-mini",
  messages: [{ role: "user", content: "Hello!" }],
});
// completion.id ↔ X-Abto-Request-Id  → link conversions to this call

Point your existing OpenAI client at our base URL. Anthropic models route through the same endpoint — just change the model name (e.g. claude-sonnet-4-5).

2. Track user outcomes

public key

Capture conversions (signup, purchase, thumbs-up…) in the browser or backend and link them to AI calls via requestId. It uses your pk-abto-… key — safe to ship in client bundles. Create a public key →

import { analytics } from "@abto/analytics-js";

analytics.init("YOUR_PUBLIC_KEY", {
  apiHost: "https://api.abto.app/v1",
});

// Page views are auto-captured. Identify and track conversions:
analytics.identify("user_123");

analytics.track("checkout_completed", {
  requestId,                 // from the AI call above — links the outcome
  properties: { amount: 49000, currency: "KRW" },
});

Zero-dependency, ~7 KB. Auto-captures $pageview on route changes, and uses sendBeacon to flush on tab close.

Try it from this browser

Mints a public key, loads @abto/analytics-js right here, and fires a $pageview + demo_click at the gateway over CORS. Then open Events to see them land.