Back every spec with customer evidence.

Kortex gathers interviews, feedback threads, competitor pages and the AI chats where ideas took shape into a Gemini Notebook per feature, so specs can cite their evidence.

100,000+ users, rated 4.5 in 410 Chrome Web Store ratings.

An example Workspace for discovery, with a Reddit thread just saved to the feedback notebook.

How do product managers use Kortex?

Product managers use Kortex, a Chrome extension for Gemini Notebook (formerly NotebookLM), to collect interview notes, customer threads from Reddit and X, competitor pages and AI brainstorms in a notebook per feature or problem. They ask it for themes with citations, keep specs in synced Google Docs, and export briefs as PDF or Markdown.

Where product evidence gets lost

Good specs cite their evidence. Kortex keeps the evidence in one place.

From discovery to the review

From customer calls to specs to the review.

Run discovery in one notebook

Themes from every interview and thread, each with its quote.

  1. Create a notebook for the problem you’re exploring, in a Discovery collection.
  2. Add interview transcripts, and send Reddit posts and X threads where customers describe the problem.
  3. Tag sources by segment or interview round.
  4. Ask for the top themes with quotes. Every point cites its source.

Track your competitors

Their pricing and docs in a notebook, with a weekly question on the calendar.

  1. Crawl each competitor’s pricing and docs pages into a notebook with the web crawler.
  2. With Kortex MCP, say once: “Every Monday at 8, ask my Competitors notebook what changed and make it an Audio Overview.”
  3. Run the market research questions prompt for the gaps worth a closer look.

Write the spec with its evidence

A spec where every claim points at a customer.

  1. Draft the PRD in Google Docs and send it to the feature’s notebook.
  2. Ask which claims in the spec have evidence in the interviews, and which don’t.
  3. As the spec changes, click Sync Google Docs to pull the latest version.
  4. Export the evidence summary as PDF or Markdown for the review.

Hand the notebook to your AI assistant

Tagging and briefs done by asking.

  1. Connect Kortex MCP to Claude, ChatGPT or Cursor.
  2. Ask: “Read the summary of every source in my Onboarding notebook and tag the ones about pricing.”
  3. Or: “Generate a briefing doc from my Customer interviews notebook and export it.”
  4. Every call runs through your own Chrome extension. See the MCP docs.

What reviewers say about these features

Quoted word for word and linked to where they were posted. Reviewers don’t say what they do for a living, so we picked reviews about the features on this page.

“Kortex solves such a practical problem. NotebookLM is powerful, but managing sources across multiple projects can get messy fast. Having a centralized way to organize, import, and export those sources with a click is exactly the kind of workflow improvement that saves hours over time.”

Product Hunt commenter, on Product Hunt

Prompts, tools and reading

What product managers ask

Something we missed? Ask us.

Can Kortex import from Jira, Linear or Intercom?

Not directly. Kortex sends web pages you can open in Chrome, Google Docs, AI chats and posts from X, Reddit and LinkedIn. For other tools, export the data as a document and add it in Gemini Notebook, or have an AI assistant with access hand it to Kortex MCP.

Is customer data safe?

Kortex doesn’t keep your chats or the contents of your sources; imports go to Gemini Notebook in your own Google account. Follow your company’s policy on customer data before adding interviews or support tickets.

Can my team use the same notebooks?

Sharing happens in Gemini Notebook itself.

How much fits in one notebook?

Gemini Notebook sets a cap on sources per notebook. Split big topics by segment or quarter, and search across every notebook to find things.

Does Kortex MCP work in Cursor?

Yes. Cursor, Claude Code, Codex, ChatGPT and other MCP clients can connect; see connect your client.

Start a notebook for your next bet.

Add Kortex to Chrome and send the next customer thread you read to a notebook.

100,000+ users, rated 4.5 on the Chrome Web Store.

Talk to the Kortex team

This opens a pre-filled email in Gmail. We usually reply within a day.