Debug your WhatsApp bot using Paige execution logs

Every time your bot handles a message, any console.log, console.error, console.warn, or console.info calls in your code are captured as execution logs. You can view these logs from the Paige dashboard to understand exactly what happened during a conversation — which is especially useful when your bot behaves differently than you expected.

View logs from the dashboard

Open your project

From the Paige dashboard, select the project you want to inspect.

Go to Tools → Logs

Open the Tools tab and select Logs. The panel shows the most recent execution output for your bot, newest at the bottom, grouped under Today and Yesterday headings.

Pick which bot you're looking at

Switch between Preview and Production at the top. This is the first thing to check when a log you expect isn’t there — see below.

Filter by log level

Use the level filter to narrow results — All Levels, Error, Warn, Info, or Log. Search box next to it matches on message text.

Preview vs Production — which bot am I reading?

You have two bots running, and they keep separate logs. The toggle at the top of the panel decides which set you’re reading, and it defaults to Preview.

Preview is the bot you’re testing — what runs when you chat in the preview panel or message through Paige Dev.
Production is the live bot your customers are talking to on your own WhatsApp number.

If a customer reports a problem and you’re staring at an empty log, this is almost always why: you’re reading Preview and the failure happened in Production. Switching reloads the panel from scratch.

Turn on Auto: On to poll for new entries every few seconds. Handy when you’re sending yourself a test message and want to watch the output land without clicking Refresh.

Log levels

Paige captures four log levels, matching the standard browser console methods:

log — General output

Use console.log() for general debugging information: variable values, which branch of logic ran, or any output that helps you trace execution flow.

console.log("Received message:", message.body);
console.log("User state:", userState);

error — Errors and failures

Use console.error() to record errors. These are shown prominently in the logs panel and are the first place to look when your bot stops responding correctly.

try {
  const result = await fetchData(userId);
} catch (err) {
  console.error("Failed to fetch user data:", err.message);
}

warn — Warnings and unexpected conditions

Use console.warn() for situations that aren’t failures but signal something unexpected — for example, a missing optional field or a fallback path being triggered.

if (!user.phoneNumber) {
  console.warn("User record missing phone number, using fallback");
}

info — Informational milestones

Use console.info() for high-level events like a flow completing, a payment being confirmed, or a user reaching a specific step.

console.info("Booking confirmed for user:", userId);

Add useful logging to your bot code

Good logging makes debugging much faster. A few practices that help:

Log inputs at the start of a handler — record the incoming message text and any user state you retrieve, so you can see exactly what your code was working with.
Log before and after external calls — if your bot calls an API or queries the database, log the request and response so you can spot failures quickly.
Log which branch ran — in conditional logic, add a short log to each branch so you can confirm the right path was taken.
Include identifiers — log user IDs or conversation IDs alongside messages so you can correlate log entries with specific users.

console.log("Handling message from:", from);
console.log("Current step:", conversationState.step);

const order = await db.orders.findOne({ userId: from });
if (!order) {
  console.warn("No order found for user:", from);
  return reply("We couldn't find your order. Please try again.");
}

console.log("Order found:", order.id, "status:", order.status);

Common error patterns

Unhandled promise rejections

If your bot makes async calls without try/catch, an error in that call will silently fail. Wrap async operations and log the error to surface the cause.

Undefined variables

A TypeError: Cannot read properties of undefined in the logs usually means a database record or API response wasn’t what you expected. Log the value before you access it.

Unexpected message types

Users can send images, voice notes, and reactions — not just text. If your bot only handles text, log the incoming message type so you can see when non-text messages arrive.

Infinite loops or repeated triggers

If the same log entry appears many times in rapid succession, your bot may be re-triggering itself. Check your reply logic for conditions that might cause a loop.

Let the Code agent read your logs

You don’t have to diagnose things from the logs yourself. The Code agent reads these same entries — just describe the symptom in chat and it will go and look:

“Customers say the booking confirmation never arrives. Check the logs.”

By default it searches both Preview and Production and tells you which one each line came from, which is often enough to establish whether a bug is live or only in your test copy. It’s the fastest route from “something’s broken” to a fix, because it can read the logs and change the code in the same conversation.

A single run records at most 500 log entries. If your code is unusually chatty — logging inside a loop over every contact, say — you’ll see Log limit reached (500). Remaining logs truncated. and the rest of that run’s output is dropped. Individual messages are cut off at 4,000 characters.

Use Paige Dev while reviewing logs — link your phone, send a test message, and watch the log entries appear. This gives you a direct line between your action and the output.

​View logs from the dashboard

​Preview vs Production — which bot am I reading?

​Log levels

​Add useful logging to your bot code

​Common error patterns

Unhandled promise rejections

Undefined variables

Unexpected message types

Infinite loops or repeated triggers

​Let the Code agent read your logs

View logs from the dashboard

Preview vs Production — which bot am I reading?

Log levels

Add useful logging to your bot code

Common error patterns

Let the Code agent read your logs