Introduction

Track what users do. Let LLM agents query the data, find insights, and tell you what to fix. No dashboards — your agent is the analyst.

LLM agents are genuinely good at analyst work now. They can segment users, spot drop-offs, correlate trends with deploys, and prescribe actions — across every metric, every morning, before you open your laptop. They just need data in a format they can read. WireLog gives them that: events in via HTTP, Markdown tables out via a pipe DSL.

10M events free. $5 per million after that. No per-seat pricing. No “contact sales.”

Drop in a script tag

One line. Automatic page_view tracking, session management, device identity, SPA navigation, UTM attribution. Done.

<script
  src="https://cdn.wirelog.ai/public/wirelog.js"
  data-key="pk_YOUR_PUBLIC_KEY"
></script>

You immediately get page views, sessions, landing pages, referrers, geo, browser/device breakdown, and navigation paths — all queryable. See what queries work out of the box.

Track custom events

Use the TypeScript client for typed events in your app. It shares identity with the script tag — same device_id, same user_id.

import { wl } from "wirelog";

wl.init({ apiKey: "pk_YOUR_PUBLIC_KEY" });

wl.track({ event_type: "signup", event_properties: { plan: "pro" } });

wl.identify({ user_id: "alice@acme.org", user_properties: { plan: "pro" } });

npm install wirelog

Server-side works too — TypeScript, Python, or Go. Pass user_id explicitly:

import { WireLog } from "wirelog";

const client = new WireLog({ apiKey: "sk_YOUR_SECRET_KEY" });

await client.track({
  event_type: "invoice.paid",
  user_id: "u_123",
  event_properties: { amount: 99 },
});

client := wirelog.New(wirelog.Config{APIKey: "sk_YOUR_SECRET_KEY"})
defer client.Close()

client.Track(wirelog.Event{
    EventType:       "invoice.paid",
    UserID:          "u_123",
    EventProperties: map[string]any{"amount": 99},
})

Query with pipes

Queries are source-first, pipe-composed. Your agent writes these — or you do, in a curl call.

# Daily signups this week
signup | last 7d | count by day

# Top pages by traffic
page_view | last 30d | count by _path | top 20

# Signup-to-purchase funnel
funnel signup -> activation -> purchase | last 30d

# Weekly retention from signup
retention signup | last 90d

# Where are users coming from?
sessions | last 30d | count by session.utm_source

# Navigation flow through your site
paths from page_view | last 30d | by _path

Output is Markdown (for agents), JSON, or CSV. Full reference in Query Language.

Let agents do the analysis

Point an agent at your WireLog project. It discovers what events exist, writes queries, reads Markdown results, and reports back with recommendations. One query at a time, across every segment, every metric.

Agent: What events are being tracked?
Query: * | last 30d | count by event_type | top 20

Agent: Signups dropped 30% this week. Let me check by platform.
Query: signup | last 14d | count by day, _platform

Agent: Mobile web signup is down. Checking the funnel.
Query: funnel page_view -> signup -> activation | last 14d | by _platform

Agent: Mobile web drops off between page_view and signup.
       The signup form may be broken on mobile. Check the deploy
       from Feb 18 -- that's when the drop started.

No dashboard can do this. An agent can do it across every metric, every morning.

Quickstart

Track events in 2 minutes. Start here

Script Tag Queries

Queries that work immediately with just the script tag. See queries

For Agents

Claude Code skills, agent patterns. Agent setup

Pricing

10M events free. $5/million after. See pricing