CommentLake — YouTube Comment Exporter

Your best product feedback is sitting in YouTube comments right now. You will not find it tomorrow. A hit video pulls thousands of comments in a week — real reactions, feature requests, buying signals, questions your audience never emailed you. Then the feed moves on and that thread disappears under a wall of new ones. This does not have to be the tradeoff. WHY PULLING YOUTUBE COMMENTS IS HARDER THAN IT SHOULD BE YouTube gives you no useful export. Studio lets you moderate, not download. The Data API has daily quotas and pagination that breaks on threaded replies. Every third-party tool adds the same friction: - Sign up for an account you will forget about - Hit a 14-day trial wall on day three of your project - Pay a monthly fee to download data you already have permission to read - Watermarked CSVs or a 100-row sample before you pay So people give up and scroll manually, screenshot comments they will never revisit, or work off Studio's small top-comment preview. The data stays trapped on the platform. This extension is the download button. WHAT COMMENTLAKE DOES CommentLake opens a side panel next to any YouTube video page and pulls the full comment thread — top-level comments and nested replies — into a clean CSV or JSON file on your machine. No sign-up. No quota counter. No watermark. You do not need API keys. You do not need a spreadsheet template. You do not need to paste a URL into a third-party site. Here is what you get: Per-comment capture — author name, author channel URL, full comment text, like count, the "published time" string (for example "2 years ago"), a direct permalink to the exact comment, type (comment or reply), parent author, and nesting depth. Video metadata — video ID, title, channel name, channel URL, and extraction timestamp. All in the same file, so a single row tells you where it came from. Two export formats — CSV for spreadsheets and BI tools, JSON for pandas, DuckDB, R, or LLM pipelines. The CSV metadata header is optional. Configurable reply depth — pick 0 for top-level only, or 1 through 5 for nested threads. Cap replies per comment so one viral thread does not dominate the export. Three extraction speeds — Fast, Normal, or Slow. Slow is the most reliable on a weak connection. Fast cuts runtime roughly in half on smaller threads. Batch mode — paste a list of video URLs, one per line. The extension runs them back-to-back in the same tab and hands you one combined file, or one file per video if you prefer. A hard cap of 100,000 comments per video — well past what any single video analysis realistically needs. Cancel anytime — mid-run cancellation saves whatever was captured, not an empty file. Eight UI languages — English, Spanish, French, German, Portuguese, Japanese, Korean, and Arabic with automatic right-to-left layout. Comment text is preserved in whatever language the user posted. HOW IT WORKS 1. Install the extension and pin the icon to your toolbar. 2. Open a YouTube video. Click the CommentLake icon — the side panel slides in. 3. Pick a comment count, a reply depth, and a format (CSV or JSON). 4. Click Extract. The panel shows live progress as comments load. 5. The file saves to your Downloads folder with the video ID in the filename. That is the whole flow. No page leaves your browser. No cap on your third run of the day. WHY A SIDE PANEL INSTEAD OF A POPUP A popup closes the second you click outside of it. Mid-extraction, that kills the progress log and can orphan a partial capture. A side panel stays open while you scroll the comments area, switch tabs for a minute, or queue up the next URL. It also has room for a readable status line and the batch queue — something a 300-pixel popup does not. You keep working on YouTube. The panel sits there doing its job. TIMESTAMPS AND PERMALINKS Every row carries the comment permalink — a URL that opens that exact comment even after the thread grows. Useful when you want to reply later or cite a quote in a report without re-finding it. The published-time string is kept as rendered ("3 days ago", "2 years ago") rather than converted to a date. That preserves what the commenter saw on the video at the time. STRUCTURED JSON OUTPUT JSON exports follow the shape { video, stats, comments }. Drops straight into pandas, DuckDB, or LLM prompt builders without reshaping. Batch output nests the same shape in a videos array. CSV INJECTION SAFETY Cells that start with =, +, -, or @ get a leading apostrophe automatically. Opening the file in Excel or Google Sheets will not silently execute a formula. REAL SCENARIOS WHERE THIS HELPS Creator sentiment audit: Your last upload has 3,400 comments. Export as CSV, sort by likes, filter on a product name. Answer in ten minutes instead of three hours. Competitive research: Paste 20 URLs into Batch mode. Come back to a combined JSON with 50,000+ comments. Run a keyword frequency pass in Python. Now you know what that audience actually talks about. Academic research: The IRB requires a verifiable source URL per quote. The permalink column gives you exactly that, per row, with the author handle intact. LLM dataset curation: Depth 2 extraction produces comment → reply → sub-reply tuples with parent author preserved. The JSON output drops into a chat-format converter as-is. SEO content planning: Pull comments from the top five videos ranking for your next topic. Those questions are the H2s your next script needs to answer. WHAT COMMENTLAKE DOES NOT DO It does not sign you in, read your account, or track you. It does not send any comment, video, or metadata to an external server. It does not post, reply, like, subscribe, delete, or modify anything — the extraction is read-only. Permissions used: - activeTab and scripting — inject the extractor into the open video page on click - sidePanel — show the control panel beside the page - storage — remember your local settings (speed, depth, format) - tabs — navigate the active tab through UR…