How to download YouTube comments and replies to CSV or JSON

YouTube Video Transcript downloads public comments and replies from a video, playlist, or channel into CSV and JSON files you can search, compare, and share. When your research spans several uploads, copying threads by hand means collecting the same details over and over. A background collection keeps the comments, their authors, and source links together, ready for your spreadsheet or research report.

Start in the Comments tool with a Google sign-in and 30 free shared credits every 30 days. No script, API key, or transcript purchase is needed. You choose the amount, whether to collect replies, and a spending limit before it runs. Each nonempty page costs 1 credit and usually contains about 20 top-level comments; reply page sizes vary. Once the job is accepted, you can close the browser and return to History for the download.

What can you do with the exported comments?

A comments export gives you feedback you can work through across videos, with a source link for each quotation. YouTube Video Transcript collects the text and its context; you decide which questions, complaints, or suggestions matter to your work.

For a product-review series, you could filter the CSV for mentions of battery life, then compare what people ask under different reviews. For a tutorial playlist, you could group questions about the same confusing step and use them to plan a follow-up video. Those are ways to use the downloaded rows in your spreadsheet, not automatic analysis performed by the collector.

When you share a finding, include the comment's link and the relevant replies. A colleague can inspect the original discussion rather than rely on an unattributed screenshot. You keep both the evidence and the route back to it.

How do you download comments from one video?

Paste the video URL into Comments, choose an order and an amount, then start a collection with a credit limit. You do not need to fetch the video's transcript first. The form separates how many comments you want from how much you are willing to spend.

1. Open Comments and paste the video link into Video, playlist, or channel URL.
2. Choose Top comments or Newest first under Order.
3. Enter a positive whole number under Amount per video, such as 20, 50, or 100. Select All available if you do not want a per-video limit.
4. Enable Include replies to collected comments if the conversations under those comments matter to your work.
5. Enter a Credit limit within your available balance and press Collect comments. The app takes you to History to follow the job.
6. When the job ends, use its download button to save the ZIP. Both CSV and JSON are included; there is no need to run it again for another format.

Your first export is usable on its own: open the CSV, find a question worth answering, and follow its source link. When you need more coverage, the same form accepts a playlist or channel. You keep the same columns and file layout as the collection grows.

Should you choose Top comments or Newest first?

Choose Top to read the comments YouTube places first, or Newest to collect recent additions. These are the two orderings described in YouTube's comments guide. The selected order is saved with your collection settings.

Top is useful when you want to inspect prominent reactions to a review or announcement. It follows YouTube's ranking, not a strict sort by likes. Newest fits a different question: what are people adding now? An older product review may attract new questions from people buying it second-hand, even when its most prominent comments concern the original release.

Label the order when sharing your findings. For example, describe a sample as “50 Top comments” so the reader knows whose feedback you examined. Neither ordering is a random audience survey.

How do amounts and replies work together?

Amount per video counts top-level comments, not every row in the export. Replies are additional and belong to the comments you actually collect. This lets you ask for 50 starting points without cutting off their conversations merely because you reached 50 rows.

As an illustration, collecting 50 top-level comments and 80 replies would produce 130 rows. Replies use the same job budget, so you control spending even when a discussion grows beyond the selected number of top-level comments.

Leave replies off when you want a first pass over questions or feedback across several uploads. Turn them on when a correction, argument, or creator response could change how you interpret a comment. A popular top-level comment can have a thread that disagrees with it; the reply count alone cannot tell you which way that discussion goes.

The export preserves is_reply and parent_comment_id, so you can connect each reply to its top-level comment and read a quotation alongside the responses it received. For interpreting those exchanges in a chat, see the separate guide to analyzing YouTube comments.

Can you collect comments from a playlist or channel?

Yes. One playlist URL, channel URL, or handle starts a collection across multiple videos. YouTube Video Transcript applies your settings to each video it reaches and packages the saved comments into one ZIP. That turns a series of video-by-video exports into a single background job you can leave running while you work on something else.

A playlist gives you a defined set of videos, such as a product-review series. A channel covers its long-form uploads, excluding Shorts. Collections support sources of up to 5,000 videos per job. For a larger source, choose a smaller playlist; the collector rejects sources above that cap.

For comparison across a series, choose a bounded amount per video and leave replies off on the first pass. For a close reading of one discussion, choose that video and include replies. You can change the scope of your research without changing tools.

The budget covers the whole job, not each video separately. Videos and their replies are collected in sequence, so an earlier discussion can use the budget before later videos are reached. You can have one active comments collection per account; finish or cancel it before starting another.

What does All available actually collect?

All available lets YouTube Video Transcript keep collecting public top-level comments beyond a fixed sample, within your credit limit. Enable replies separately to collect their threads too. This is the setting for a broader dataset when the first page is not enough.

Availability still matters: the collector cannot recover deleted comments or access private discussions, and comments must be enabled on the video. YouTube explains that authors, channel owners, and moderation can remove comments in its comments help page.

Collection ends when pages run out, the budget is used, or you cancel. A failed fetch can also leave partial results for a video. The ZIP's manifest.json records settings, credit use, any stopping message, and each discovered video's status, so you can see the coverage you received. Check it before calling a dataset complete: completed can also mean the job reached its budget and saved a partial collection.

What is inside the CSV and JSON download?

Every video with saved comments gets a CSV file and a JSON file in the same ZIP. Both contain the collected top-level comments and replies. A separate manifest describes the job and its coverage. For one video with saved comments, the file layout is below; VIDEO_ID is a placeholder.

manifest.json
VIDEO_ID.comments.csv
VIDEO_ID.comments.json

Open the CSV to review feedback in a spreadsheet, or pass the JSON to your own analysis code. There is no second collection to run and no second format to pay for. For a video with no saved comments, the manifest records its status instead of an empty comments file.

Likes and publication times are display text, not guaranteed exact numbers or calendar dates. A value such as 3K should not be treated as an exact count of 3,000. A parent's reply count describes the discussion reported by YouTube; count the saved reply rows to learn how much of it is actually in your file.

When combining files from several videos, add the video ID from each filename as a column. Keep the manifest alongside the combined file so that collection settings do not get separated from the data. The comment links let you check quotations while the original comments remain available on YouTube.

How much does a comments collection cost?

You choose the maximum spend before collection starts. YouTube Video Transcript charges 1 shared credit per nonempty page, including reply pages; empty pages and failed requests cost nothing. Downloading the saved ZIP and using either format adds no charge.

Top-level pages usually contain about 20 comments; reply pages can be smaller. A five-credit limit therefore allows at most five charged pages, not a guaranteed 100 rows. Your requested amount may also keep only part of the last page. Set the limit for your budget and use the exported row count to describe the result.

The limit is reserved when you submit, and unused credits are released when the job ends within the same billing cycle. For example, reserve 10 and use 3, and the unused 7 return to that cycle's balance. You can cancel from History and keep the comments already saved. Reservations belong to the cycle that funded them; an old-cycle refund does not add credits to a new cycle. Full rules are in the website guide.

Free accounts can use the collector within their 30 credits every 30 days. Paid plans provide a larger shared balance: Starter is $9 a month for 1,000 credits, Pro is $19 for 5,000, and Business is $49 for 20,000. No separate comments subscription is required. See pricing for the plans; unused credits do not roll over.

Which comments workflow should you use?

Choose YouTube Video Transcript's standalone collector for a downloadable dataset with a budget you control. If you need a quick read or context beside a transcript, the other workflows below let you keep the task that small.

• For downloadable comments across a video, playlist, or channel, use YouTube Video Transcript's Comments tool. Choose the amount, replies, and budget before it runs.
• For a handful of comments you only want to read, open the video on YouTube. No export or credit balance is needed to read the public discussion there.
• For context beside a transcript, the Workbench offers a separate Top or Newest sample, usually about 20 comments, for 1 credit when nonempty. Transcript bulk jobs can include a Top sample per successful transcript. Use the standalone collector for more pages or replies.
• For questions inside an AI chat, use the comments-analysis workflow. MCP fetches pages for the conversation; it does not start a standalone comments collection job.
• For your own collection code, our comments API exposes pages and reply continuations. The official YouTube Data API is another source of comment threads for a script. In that workflow, you handle collection and file creation yourself.

Questions before your first export

A comments-only collection does not depend on captions, and you can leave the browser once the job has been accepted. These are the remaining details to check before starting.

Can I download YouTube comments without downloading a transcript?

Yes. YouTube Video Transcript's Comments tool collects public comments independently of captions. Paste a video, playlist, or channel URL, choose your collection settings, and set a credit limit. You do not need to buy a transcript, and videos without captions can still supply comments when those comments are publicly available.

Does the amount per video include replies?

No. Amount per video limits top-level comments. Replies to those collected comments are additional when Include replies is enabled. Both top-level and reply pages use the same job credit limit, so the final export can contain more rows than the selected amount without exceeding that budget.

Can I export all comments from a YouTube channel?

YouTube Video Transcript accepts a channel URL or handle and can collect comments across its long-form videos. Select All available to remove the per-video comment limit. The job still respects your credit limit, public availability, and the maximum source size of 5,000 videos. Channel collections exclude Shorts; sources larger than the cap need a smaller playlist. Check the export's manifest before describing it as complete.

How many comments does one credit buy?

One credit pays for one nonempty comments page, including a reply page. Top-level pages usually contain about 20 comments, but page sizes vary and reply pages can be smaller. It is not a fixed price per 20 exported rows. Empty pages and failed requests cost no credits, and downloading a saved ZIP costs nothing extra.

Can I close the browser or cancel the collection?

Yes. Once YouTube Video Transcript accepts the job, collection continues in the background. Return to History to download it, or cancel an active job there. Cancellation keeps saved comments and releases the unused reservation within the billing cycle that funded it. Download links expire 100 days after the job ends, so save the ZIP locally if you need to keep it longer.

Make your first comments export with free credits

Sign in with Google to get 30 free shared credits every 30 days. Open Comments, paste a video you want to research, and set your collection amount and spending limit. You can try the full CSV and JSON export on the free plan, including replies within your budget, without buying a transcript.

Use the first file to answer a real question: what keeps coming up, which questions need a response, or what viewers want covered next. When you need to compare more uploads, give YouTube Video Transcript a playlist or channel and collect the feedback in one job.