Now in beta · Powered by Gemini 2.5 & GPT-4o

Debug at the speed of thought.

Lumen is an AI assistant built into Chrome DevTools. It sees what you see, answers what you ask, and runs code directly on your page.

Get started free Watch 30s demo
6 AI modelsin one extension
<300msto first token
10,000+ requestsserved daily
Why Lumen

Built for the way you actually debug.

Six tools that turn DevTools from a console into a conversation. The browser stays open. The flow stays unbroken.

Unique to Lumen

Run code on the page.

Lumen writes a fix. You click Run. The code executes on the live page and returns the result inline. No copy-paste. No console. No tab switching.

chrome.devtools.inspectedWindow.evalfull page scope
02

Inspect any element

Select a button in Elements. Lumen reads its HTML, computed styles, layout, and event listeners — then answers your question about it specifically.

computed styles · event listeners
03

Explain errors instantly

Every console.error, uncaught exception, and rejected promise is captured. Click to get a diagnosis and a fix in 3 seconds.

diagnosis in ~3s
04

Six AI models, one click.

Gemini 2.5 Pro for hard problems. Flash for speed. GPT-4o when you need OpenAI's reasoning. Switch mid-conversation.

Gemini · GPT-4o · o4-mini
05

Streaming responses

Real SSE streaming. Words appear as the model thinks. No spinning, no waiting for the full response.

<300ms to first token

Built for privacy.

Chat history stays in your browser. Page context is sent per request, never stored. Your code never trains anyone's model.

local-firstzero trainingper-request context
The killer move

From "what's wrong" to "fixed" in one click.

Most AI tools hand you a snippet and walk away. Lumen ships the fix. Generated code executes in the same JavaScript context as the page, gives you back the return value, and lets you keep moving.

  • Runs inside chrome.devtools.inspectedWindow.eval — full page scope.
  • Return value streams back into the chat — including DOM nodes and objects.
  • Confirmation modal before anything destructive.
How it works

Three steps to a smarter DevTools.

01

Install the extension.

Add Lumen from the Chrome Web Store. 30 seconds.

02

Open DevTools on any page.

F12 → find the Lumen tab. Sign in.

03

Ask anything.

Attach an element. Paste an error. Lumen answers — and can execute the fix.

Pricing

Credits, not seats.

Pay for what you use. Different models cost different credits. Unused credits roll over.

Free
$0forever
For curious developers.
  • 100 credits / month
  • All 6 AI models
  • Run on page
  • Chat history
Starter
$4.99/mo
For daily debugging.
  • 500 credits / month
  • Everything in Free
  • Rolls over each month
  • Priority support
Most popular
Pro
$9.99/mo
For full-time builders.
  • 1,500 credits / month
  • Everything in Starter
  • Premium models prioritized
  • Rolls over (cap 3×)
Max
$19.99/mo
For teams that ship daily.
  • 5,000 credits / month
  • Everything in Pro
  • Team support coming soon
  • Highest rate limits
Credit pack add-on
+1,000 extra credits · never expire · one-time purchase
$7.99
Compare

One column gets a checkmark in every row.

Lumen ChatGPT GitHub Copilot
Lives in DevTools✓——
Sees inspected element✓——
Runs code on the page✓——
Captures console errors✓——
Multiple AI models✓——
In-browser, no IDE✓Partial—
Pricefrom $0$20/mo$10/mo
FAQ

The questions developers actually ask.

What does "Run on page" actually do?
When Lumen returns a JavaScript snippet, click the green Run button. The code executes in the inspected page's JavaScript context (the same as if you typed it in the DevTools console). The return value appears inline in the chat. This works because Lumen is a Chrome DevTools extension with access to chrome.devtools.inspectedWindow.eval.
Which browsers are supported?
Chrome, Edge, Brave, Arc, and any Chromium-based browser. Firefox is on the roadmap.
Can my team use one account?
Each account is tied to one device. Need multiple seats? Reach out and we'll sort you out.
Is my browsing data sent anywhere?
Page context (element HTML, error messages) is sent only when you explicitly attach it, and only to the AI model for that single request. We don't store it. Chat history stays in your browser.
Can I cancel anytime?
Yes. Cancel from your profile. Pro access continues until the end of the billing period.

Stop tab-switching. Start shipping.

Lumen runs free, forever. Install in 30 seconds.

Install free