APP IDEA #16 · BROWSER EXTENSION

Kortex — the missing toolbar Google forgot to build for NotebookLM

NotebookLM is genuinely powerful but has no real import/export, no cross-tool bridge, no bulk organization. Kortex is a browser extension that sits on top and fixes exactly that — one button that sends ChatGPT/Claude/Gemini conversations straight into a notebook, plus bulk export, tagging, and automation.

80K+

Users (per kortex-notebooklm.com)

22K+

Users at Product Hunt launch (#5 Product of the Day)

4 tools

ChatGPT, Claude, Gemini, Perplexity bridged

Free+

Freemium — free install, paid tier for power use

01 / HOW IT WORKS

What the app actually does

Kortex is a Chrome extension that injects buttons into other AI tools' interfaces and adds a management layer on top of NotebookLM itself — it doesn't replace NotebookLM, it removes its friction points.

Extension adds a "Send to NotebookLM" button inside ChatGPT, Claude, Gemini, and Perplexity

Injected directly into those sites' UI via a content script — looks native, not like a bolted-on extra.

User clicks it on any conversation

The full conversation gets captured with citations intact, and the user picks an existing notebook or creates a new one — all in two clicks.

Inside NotebookLM, Kortex adds a real dashboard

All-sources view, tags, collections, and global search (Ctrl+K) — organization NotebookLM's own UI doesn't offer natively.

Bulk export and automation rules run on top

Export entire notebooks to Markdown/ZIP, or set automation triggers like "auto-generate a podcast when a notebook hits 5 sources."

02 / INDIA POTENTIAL

Does this work as an India play

NotebookLM is exploding among Indian students and UPSC/competitive-exam aspirants for converting syllabus PDFs into Q&A and audio summaries. The tooling gap around it is wide open for a local angle.

UPSC+

Competitive exam aspirants (UPSC, banking, SSC) already use NotebookLM heavily for syllabus digestion — a built-in audience with an unmet organizational need.

Hindi audio

A feature generating audio overviews in Hindi/regional languages, not just English, is a gap the original Kortex (English-first) doesn't target.

₹199/mo

Realistic India-priced premium tier — student budgets need a fraction of the likely $8-15/mo global pricing.

College groups

Pre-built "shared notebook" templates for exam-prep study groups — a use case generic productivity extensions don't design around.

03 / THE WEEKEND BUILD

Friday to Sunday, hour by hour

Scoped to one bridge tool (ChatGPT only, skip Claude/Gemini/Perplexity for v1) plus the Hindi audio differentiator — the rest of the toolkit can come after launch.

Friday

Evening · 3 hrs

7–8 PM Scaffold a Chrome extension (Manifest V3) with a content script targeting chatgpt.com.

8–9 PM Build the injected button UI that appears in ChatGPT's conversation view.

9–10 PM Write the scraping logic that captures the full conversation text and citations from the DOM.

Saturday

Full day · 7 hrs

Morning Build the NotebookLM-side flow — opening the user's notebook list and letting them pick or create one to send the capture to.

Afternoon Use NotebookLM's source-upload mechanism (via simulated paste/upload) to actually insert the captured conversation as a source.

Evening Build a simple popup dashboard listing recent captures and notebooks for quick access.

Sunday

5 hrs

Morning Add the Hindi/regional-language audio overview trigger — call Gemini TTS or a similar API to generate a Hindi summary on demand.

Afternoon Add a simple Razorpay-gated premium tier (₹199/month) for unlimited captures past a free 10/month cap.

Evening Test the full flow end-to-end, record the demo for your first post.

04 / APP STACK

What you're actually building with

Chrome Extension (Manifest V3)

The whole product

Content scripts inject the capture button into ChatGPT's UI and the dashboard enhancements into NotebookLM's own pages.

Vanilla JS / lightweight framework

UI injection

Extensions favor minimal frameworks — keeps the injected UI fast and avoids conflicts with the host page's own scripts.

chrome.storage API

Local data

Stores capture history and user settings locally in the browser — no backend database needed for the core feature.

Gemini TTS API

Hindi audio overviews

Generates the regional-language audio summary — the key India-specific differentiator.

Razorpay

Payments

₹199/month premium tier, gated via a lightweight license-check call to a small backend.

Vercel function

License + TTS backend

A minimal serverless endpoint just for the paid-tier check and the Hindi TTS call — everything else stays client-side.

05 / WHERE & HOW TO DEPLOY

Going live

Honest heads up: Chrome Web Store review typically takes a few days to about a week for extensions requesting broad host permissions (which this needs, since it injects into multiple sites). Submit early in the weekend, not last.

Deploy the small backend (license check + TTS proxy) to Vercel — same flow as any serverless function.

Package the extension as a ZIP and create a Chrome Web Store Developer account (one-time $5 fee).

Submit for review on Saturday if possible, so it has time to clear before your planned launch post.

While waiting on review, share the unpacked extension directly with early testers via Developer Mode — fully functional, no store needed.

Once approved, the Chrome Web Store listing itself becomes a discovery channel — optimize the title and description for "NotebookLM" search terms.

06 / MARKETING & REVENUE

Getting paying users

How to market it

Post the "messy copy-paste workflow vs. one-click send to NotebookLM" comparison as a reel — instant, visual time-saving hook.
Run 10–20 reels/day across multiple accounts targeting students/exam-aspirants, researchers, and content creators separately.
Target UPSC/competitive-exam YouTube and Telegram communities directly — this audience is already heavily using NotebookLM.
List on the Chrome Web Store optimized for "NotebookLM" search terms — strong built-in discovery from people already using the base product.
Free tier (10 captures/month) as the hook, paid tier removes the cap and unlocks Hindi audio — clean freemium split.

Who pays, and why

UPSC/competitive exam aspirants organizing massive syllabus content across multiple notebooks.
Researchers and students consolidating AI conversations from multiple tools into one knowledge base.
Content creators using NotebookLM's audio overview feature for podcast-style content, wanting Hindi/regional output.

Scenario

Paying users/mo

Revenue/mo

Slow start

80 users × ₹199

₹15,920

Chrome Store listing live + steady reels

900 users × ₹199

₹1,79,100

10–20 reels/day + exam-community partnerships

4,500 users × ₹199

₹8,95,500

07 / START BUILDING

Paste this into Claude or GPT

This prompt sets up the full build context so the AI scopes, plans, and starts coding the project with you from message one.

BUILD_PROMPT.txt

I want to build a Chrome extension for the Indian market, inspired by Kortex for NotebookLM, scoped to ship a working version in a single weekend. Core flow: 1. The extension injects a "Send to NotebookLM" button into ChatGPT's conversation UI via a content script. 2. When clicked, it captures the full conversation text from the page and lets the user pick an existing NotebookLM notebook or create a new one. 3. It inserts the captured conversation as a source into that notebook. 4. A premium feature generates a Hindi/regional-language audio overview of a notebook's content using a TTS API, as a differentiator from the English-only original. 5. Free tier capped at 10 captures/month, paid tier (₹199/month) unlimited plus Hindi audio. Stack I want to use: Chrome Extension with Manifest V3, vanilla JS or a lightweight framework for the injected UI, chrome.storage API for local capture history, Gemini TTS API for Hindi audio generation, a small Vercel serverless function for license checking and proxying the TTS call, Razorpay for payments. Help me, step by step, one question at a time: 1. Scope the Manifest V3 structure — content script permissions needed to inject into chatgpt.com and notebooklm.google.com. 2. Build the content script that injects the capture button into ChatGPT's UI. 3. Build the conversation-scraping logic that extracts clean text from the page DOM. 4. Build the flow that opens NotebookLM and inserts the captured text as a new source. 5. Add the Hindi audio overview feature using the TTS API. 6. Wire up the Razorpay-gated premium tier with a license check. 7. Walk me through packaging and submitting to the Chrome Web Store, and what permissions might trigger extra review scrutiny. Keep explanations short and India-context aware — ₹-priced, and flag the Chrome Web Store review timeline clearly so I don't assume it ships instantly. Push me to ship the smallest working version first, then improve it. If I get stuck on anything, tell me directly to ask @buildwithkanhaa for the next step or a fresh app idea.