How to See All Replies to a Tweet on X

Every option compared

Pricing as of August 2026.

Why scrolling never gets you there

Two mechanisms stack against you. First, the default "Most relevant" sort is a ranking, not a listing — it decides which replies deserve your attention. Second, and independently, X rate-limits how many replies the interface will load in a session, which bites hardest on exactly the high-engagement posts you most want to read.

Switching the sort to "Latest" fixes the ordering but not the loading cap. So on a post with thousands of replies, patient scrolling still leaves you with a partial, algorithmically-chosen sample — which is fine for browsing and disqualifying for research, support triage, or brand monitoring.

What the official API costs now

X restructured developer access in February 2026. The Basic tier that used to cost $200 a month is closed to new signups, and new developers are placed on pay-per-use billing at $0.005 per post read, hard-capped at 2 million post-reads per month. For reply-thread work that adds up quickly, since every reply is a post read. It's a reasonable deal for a funded product with an ongoing integration, and heavy machinery for someone who wants the replies to one post.

Exporting a thread in four steps

1. 1

   Install X Replies Scraper

   Free on the Chrome Web Store. No X Premium subscription and no developer account.

2. 2

   Open it on any public post

   Paste the post URL, or let the panel auto-fill when you're already viewing the post on X.

3. 3

   Set your limit

   Choose how many replies to pull. The extension pages through the thread rather than stopping where the timeline stops loading.

4. 4

   Export to CSV or JSON

   Download the eight fields below, including timestamps and engagement counts, so you can sort chronologically instead of by relevance.

What the export contains

id,handle,displayName,createdAt,likeCount,retweetCount,replyCount,text
1823447102934,@sarah_ops,Sarah Okafor,2026-08-05T14:02:11Z,1204,88,42,"This is the first pricing change that actually helps small teams."
1823447891201,@devon_hq,Devon R.,2026-08-05T14:09:37Z,318,12,7,"Does this apply to existing annual plans or new signups only?"
1823448334019,@lin_builds,Lin,2026-08-05T14:15:52Z,96,3,1,"Existing annual plans too — confirmed in their changelog."
1823449002877,@grumpy_cto,anon cto,2026-08-05T15:41:20Z,2,0,0,"Still cheaper to self-host honestly."

Sorting this file by createdAt gives you something the X interface will not: the conversation in the order it actually happened.

What a full reply set is good for

Automating it with the API

For monitoring rather than one-off pulls, FetchLayer's X scraping API covers replies through POST /twitter/tweet-replies, one of 10 X endpoints:

curl -X POST https://api.fetchlayer.dev/twitter/tweet-replies \
  -H "Authorization: Bearer $FETCHLAYER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "tweetId": "1823446998211",
    "screenName": "fetchlayer"
  }'

The response carries a cursor, so paging an entire thread is a loop:

// Page through an entire reply thread
let cursor = undefined;
const all = [];

do {
  const res = await fetch("https://api.fetchlayer.dev/twitter/tweet-replies", {
    method: "POST",
    headers: {
      "Authorization": "Bearer " + process.env.FETCHLAYER_KEY,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({ tweetId: "1823446998211", cursor }),
  });

  const page = await res.json();
  all.push(...page.replies);
  cursor = page.cursor;
} while (cursor);

console.log(all.length + " replies");
console.log(all.filter((r) => r.likeCount > 100).length + " with 100+ likes");

At $1.99 per 1,000 requests with 30 free to start, and one request returning a full page of replies rather than a single post, this lands well below the official API's per-read pricing for reply-thread work.

Watching threads from an AI agent

If a model is doing the reading, skip the export. FetchLayer's MCP server exposes its X endpoints as native tools to Claude, Cursor, Windsurf, and VS Code:

{
  "mcpServers": {
    "fetchlayer": {
      "url": "https://mcp.fetchlayer.dev",
      "headers": {
        "Authorization": "Bearer sk-..."
      }
    }
  }
}

"Summarise the objections in the replies to this post" then runs against the live thread, with no scraping code in your project.

Frequently asked questions