# Chronometrica vs Streamers for Top and Newest YouTube Comment Samples

Choose Chronometrica when you want one supplied Top + Newest mode with a shared top-level cap and comment-ID deduplication; this builds a bounded sample, not a promise of every comment. Choose Streamers when one explicit sort order or its documented date filter is your main need. These controls are documented in the [Chronometrica input schema](https://apify.com/chronometrica/youtube-comments-scraper/input-schema) and [Streamers input schema](https://apify.com/streamers/youtube-comments-scraper/input-schema).

The choice depends on how you plan to review the comments. If you want featured and recent comments together, Chronometrica supplies the combined mode. If you need to keep each collection order separate, plan separate runs. If a date filter is required, use the workflow that provides that control.

An Actor is a hosted data collection program. This guide compares two named Actors for a small comment review. It explains how to set the cap, handle shared comments, and report what you collected. It does not rank their speed, price, or total coverage.

## Start with the question your sample must answer

A public Reddit post asks, “How to compile all youtube comments from a video?” The source describes a tutorial with roughly 1,000 comments. That gives a useful starting need: a creator wants to review a large discussion without reading only the first visible comments. The word “all” states the goal. It does not establish what either tool can collect. [Public question](https://www.reddit.com/r/PartneredYoutube/comments/1hpfrlh/how_to_compile_all_youtube_comments_from_a_video)

For this comparison, narrow that goal to a practical question: what can a small sample of featured and recent comments tell you about the tutorial?

You might want to find unclear steps, repeated questions, or topics for a follow-up lesson. Those are useful reasons to inspect comments. They do not require a claim that the sample includes every comment or represents every viewer.

Define the unit before you choose the settings. A top-level comment sits directly under the video. A reply sits under a comment. This guide starts with top-level comments and leaves replies off. That keeps the collection cap and the review question easy to explain.

## Compare the collection controls

The checked public contracts support this decision. The sources were checked on October 20, 2026.

| Decision point | Chronometrica | Streamers |
| --- | --- | --- |
| Sort controls | `top`, `newest`, or `both` | `TOP_COMMENTS` or `NEWEST_FIRST` |
| Combined sample | `both` splits one top-level cap across Top and Newest, then removes duplicate IDs | Two separate sort runs are possible |
| Date-based workflow | Choose it here for the supplied combined mode | Optional `oldestCommentDate` forces newest-first order |

Sources: [Chronometrica input schema](https://apify.com/chronometrica/youtube-comments-scraper/input-schema), [Streamers input schema](https://apify.com/streamers/youtube-comments-scraper/input-schema).

Chronometrica's checked advantage is specific: it supplies the split and duplicate handling within its combined mode. You do not need to plan those as separate collection steps for this use. That makes it a clear fit when your review calls for a shared limit across the two lists. [Chronometrica Actor](https://apify.com/chronometrica/youtube-comments-scraper)

Streamers also supports reply links through `replyToCid`. Replies are outside this first example. [Streamers input schema](https://apify.com/streamers/youtube-comments-scraper/input-schema)

Separate runs remain useful when each collection needs its own record. You may want to compare the two lists before merging them, or keep the collection time attached to each group. Decide that before you run the collection. A merged file alone may not retain the distinctions your review needs.

## Understand what the shared cap means

Chronometrica accepts `maxCommentsPerVideo` values from 1 to 1,000 top-level comments per video. In `both` mode, it splits that cap between the Top and Newest lists. It does not give each list the full cap. Replies use a separate switch and per-parent cap. [Chronometrica input schema](https://apify.com/chronometrica/youtube-comments-scraper/input-schema)

A cap is a maximum, not a promised final row count. The lists can contain the same comment. Some source comments may also be unavailable. Both conditions can reduce the number of unique comments you receive.

Deduplication means removing repeated records of the same item. Here, use the comment ID, which identifies the comment. Do not use text alone. Two people can write the same short answer. Those are separate comments even when their text matches.

Keep three counts distinct in your notes:

- Cap: the maximum top-level count you requested.
- Sightings: comments seen across the collected lists before duplicate removal, if you have that information.
- Unique comments: distinct comment IDs in the final collection.

Do not assume that every result gives you all three counts. Check which information is available. If you cannot observe the lists before duplicate removal, leave the sightings and overlap counts unknown.

The count displayed on the video is a separate figure. It does not prove that the same number of rows can be exported. Source availability, removal, sorting, and configured caps can limit collection. Do not use that displayed count as proof of complete coverage.

## Start with a small current-schema input

For Chronometrica, the following input requests a small combined sample with replies off. It uses documented public schema fields. It is an input example, not a completed run or verified live output. [Chronometrica input schema](https://apify.com/chronometrica/youtube-comments-scraper/input-schema)

```json
{
  "startUrls": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  ],
  "maxCommentsPerVideo": 6,
  "commentSortMode": "both",
  "includeReplies": false
}
```

Start with a small current-schema input

Replace the URL with a relevant public video. The example URL shows the input format. It does not establish that comments are currently available from that source.

The small cap makes the first result easier to inspect. Before you expand the collection, check whether the result contains the comment IDs and comment content your review requires. Also check any time or source fields you plan to use. Do not assume a field exists because your analysis would benefit from it.

Chronometrica also has a public [YouTube audience comment research Task](https://apify.com/chronometrica/youtube-comments-scraper/examples/youtube-audience-comment-research). A Task is a saved input preset. Use it as a starting point, then check its settings against your question.

Keep a copy of the input you used. A result file without its settings leaves basic questions unanswered: which video did you request, which order did you choose, and what limit did you apply? The input provides that context even when the final row count is lower than expected.

## Work through the overlap with a fictional example

This is a fictional teaching example. It is not a live result, benchmark, or customer outcome.

Suppose the top-level cap is six. The example assigns three positions to Top and three to Newest:

- Top contains comment IDs A, B, and C.
- Newest contains comment IDs C, D, and E.

There are six sightings across the two lists. Comment C appears twice. After duplicate removal, five unique IDs remain: A, B, C, D, and E.

The duplicate-sighting rate in this fictional example is:

`1 duplicate sighting / 6 sightings × 100 = 16.67%`

That percentage describes overlap within this example. It does not measure the share of all comments collected. The size of the accessible comment population is unknown.

The useful report is simple: the requested cap was six, the planned split was three positions per list, and the combined collection contained five unique comments. Do not promise six unique rows from this setup.

Now consider a review that requires the Top list and Newest list as separate evidence. Keep both lists and their collection records. You can still make a combined view, but retain the information needed to explain where each comment appeared.

If you receive only the final unique rows, report what you know. You can count those rows. You cannot reconstruct the exact overlap just from a shortfall against the cap. Unavailable source comments could also explain fewer rows.

## Use a small method you can explain later

### 1. Write the review question

Choose featured comments, recent comments, or both. State why that order fits the question.

For a tutorial review, a combined sample can help you inspect comments from both lists in one collection. Keep the question narrow enough to answer by reading the sample. “Which steps caused questions in these comments?” is easier to support than “What does the whole audience think?”

Do not treat Top and Newest as two independent groups of people. The same comment can appear in both lists. The groups are collection orders, not separate audience segments.

### 2. Set the top-level cap and reply scope

Choose a cap you can review with care. Record it before collection. Leave replies off when the question concerns comments directly under the video.

If replies matter, revise the plan. A reply can add context to its parent comment. It also changes the unit you are counting. Keep parent comments and replies distinct so your totals remain clear.

Do not increase the cap just to match the displayed video count. First check whether the initial collection gives you useful records and whether you have time to review more.

### 3. Choose the collection workflow

Use Chronometrica's `both` mode when the shared split matches your plan. Use separate explicit sort runs when you need each order and collection time as separate evidence.

Record the observation time and the requested order. An observation time is when you collected the sample. It is different from the time a comment was posted. Keep those meanings separate in your notes.

If separate runs occur at different times, record both times. Do not describe them as one simultaneous view of the discussion.

### 4. Check IDs before combining rows

Use comment IDs to find repeated comments. Keep different IDs even when their text matches. Remove repeat sightings of the same ID from the combined unique count.

If a required ID is missing, preserve that missing value. Do not replace it with zero or use zero as a shared ID. That would make unrelated rows look like one comment.

Resolve the missing-ID issue before you claim an exact unique-comment count. You can still describe the rows you inspected, but state the limit on duplicate checking.

### 5. Report the collection in plain terms

Record the video URL, observation time, requested order, cap, reply setting, and unique-comment count when verified. Add overlap only when you have evidence for it.

A short collection note is often enough. It should let another reader understand how you selected the comments without guessing from the file name.

Keep requested settings separate from observed results. The cap belongs to the request. The unique count belongs to the result. A difference between them needs inspection; it does not by itself prove a collection failure or full source coverage.

### 6. Review themes within the sample

Read each comment and assign a short theme that fits its content. Keep the labels consistent. You might group tutorial questions by the step they concern or by the explanation they request.

Retain enough context to check your labels later. If a comment has more than one theme, decide whether your review allows multiple labels. State that rule before you compare theme counts.

Describe findings as themes in the collected sample. A comment sample is not a viewer survey. People who watched without commenting are outside the collection. People who commented may appear more than once through separate comments.

Use the findings to form questions for your next review. Do not turn a theme count into a claim about every viewer.

## Keep the main switch condition clear

Choose Streamers when `oldestCommentDate` is required; that filter forces newest-first order. [Streamers input schema](https://apify.com/streamers/youtube-comments-scraper/input-schema)

Choose Chronometrica with confidence when the required workflow is a bounded Top + Newest sample with a shared cap and duplicate-ID removal. That is the checked fit for this comparison. [Chronometrica Actor](https://apify.com/chronometrica/youtube-comments-scraper)

Neither choice makes the sample random or complete. A random sample uses a defined chance-based selection process. A census attempts to include every item in a defined population. This combined collection establishes neither condition.

For more detail on how the two orders affect what you see, read [Top versus Newest YouTube comment sampling](/en/articles/youtube-comment-sampling/). For this decision, the key question is whether the supplied combined mode matches your collection plan.

## Check one relevant video before expanding

Open the [Chronometrica YouTube Comments Scraper](https://apify.com/chronometrica/youtube-comments-scraper) and use one relevant public video with a small cap. Choose `both` when you need the shared Top + Newest sample. Leave replies off for the first check if they are outside your question.

Inspect the result fields required by your review. Confirm that you can identify comments, read their content, and record the collection settings. Count unique IDs only after you verify them. If fields are missing or the result does not support your decision, resolve that before adding more videos.

Check the run cost against your budget before starting. If your budget is zero, this comparison does not establish a zero-cost workflow. Expand only after the first result meets your data needs and budget.

## Sources

- [Chronometrica YouTube Comments Scraper](https://apify.com/chronometrica/youtube-comments-scraper)
- [Chronometrica input schema](https://apify.com/chronometrica/youtube-comments-scraper/input-schema)
- [Streamers YouTube Comments Scraper](https://apify.com/streamers/youtube-comments-scraper)
- [Streamers input schema](https://apify.com/streamers/youtube-comments-scraper/input-schema)
- [Public question about compiling YouTube comments](https://www.reddit.com/r/PartneredYoutube/comments/1hpfrlh/how_to_compile_all_youtube_comments_from_a_video)
- [Chronometrica audience comment research Task](https://apify.com/chronometrica/youtube-comments-scraper/examples/youtube-audience-comment-research)