Podcast Transcript Search Overlay

WHAT IT DOES

Podcast Transcript Search Overlay is a powerful search tool that lets you instantly find any word or phrase within a podcast transcript. It automatically detects transcript content on podcast episode pages—whether it's visible text transcripts, closed captions, or embedded transcript sections—and provides a searchable overlay where you can enter keywords and see matching lines with surrounding context. Each result displays the timestamp (when available), speaker label, and the matched line highlighted in your chosen color. The tool also shows up to two lines of context before and after each match, helping you understand the full conversation around your search term. You can export all search results with timestamps to plain text or Markdown format, making it easy to reference key moments, create show notes, or share quotes from the episode.

WHERE IT RUNS

This script runs on any webpage that contains podcast transcript content—podcast episode pages, blog posts with embedded transcripts, podcast hosting platforms (Transistor, Buzzsprout, Libsyn, Anchor, Podbean, Simplecast), and podcast directories. It operates entirely within your browser tab, reading the page's DOM to locate transcript containers. The script cannot access other tabs, your browsing history, bookmarks, downloads, or any data outside the current page. All settings—including case sensitivity, results per page, highlight color, auto-open preference, and export format—are stored locally using GM_setValue and never transmitted anywhere.

HOW TO USE

After installing with the BotGentz extension, the script automatically detects transcripts on podcast episode pages. When a transcript is found, you'll see a note that the transcript is loaded and ready for searching. Enter a keyword or phrase in the search box and click the search button or press Enter. Results appear instantly, showing each matching line with surrounding context, timestamp, and speaker label (when available). Each result is highlighted with a colored border for easy scanning. Export search results to plain text or Markdown using the export button—great for creating show notes, social media posts, or research references. You can also copy the entire raw transcript with one click. Toggle case sensitivity for precise searches, adjust the highlight color, and set auto-open on transcript detection.

THE REAL MECHANISM

The script uses a comprehensive DOM scanning strategy to locate transcript content across diverse page structures. It queries a wide range of selectors including .transcript, .transcript-text, .captions, .closed-captions, .podcast-transcript, and many more. If no dedicated transcript container is found, the script scans the page for elements containing timestamp patterns like [00:00], (00:00), or HH:MM:SS, and selects the largest such block as the transcript. Once extracted, the script parses each line, extracting timestamps using regular expressions (supporting both MM:SS and HH:MM:SS formats), and detects speaker labels (e.g., "Host:", "Guest:") using pattern matching. The search algorithm performs a case-sensitive or case-insensitive match based on user preference, then builds context windows of two lines before and after each match. For each result, the script creates a visual highlight by wrapping matches in span elements with the user's chosen background color. All rendering uses safe DOM manipulation with no innerHTML from untrusted content.

THE PANEL

The Transcript Search panel is a BGPanel component that you can drag anywhere on your screen by its header. It snaps to the nearest edge when released, keeping your workspace organized. The panel collapses to a small bubble showing only the emoji icon when minimised—click to expand it again. You can resize the panel horizontally to see more of the search results. Press the Escape key to collapse the panel instantly. Your panel position and size are saved per-site.

PLEASE NOTE

This script requires the free BotGentz browser extension to function. Transcript extraction depends on the page structure—some podcast sites embed transcripts in non-standard containers that may not be detected. If the script doesn't find a transcript, use the "Extract Transcript" button to retry detection. Timestamp and speaker detection rely on common formatting patterns; custom formats may not be parsed correctly. The script never reads or writes password fields, hidden inputs, or file inputs. It stores only your case sensitivity preference, results per page, highlight color, auto-open preference, and export format locally. No transcript content, search queries, or personal information is ever collected, transmitted, or shared.