Archive your own Threads posts as JSON before deleting the account
Pull a public creator's posts and replies into Notion or Obsidian
Export a profile to CSV for analysis in a spreadsheet
Capture reply threads where the profile owner participated using Deep mode
Requires Tampermonkey installed in the browser, and only works on public profile pages because there is no login or network interception.
threads-scraper is a small browser script that downloads every post from a public profile on Threads, the social network from Meta. It is delivered as a userscript, which is a kind of mini-extension you load through a browser add-on called Tampermonkey. Once installed, you visit any Threads profile URL of the form https://www.threads.com/@username and a small panel appears in the top-right of the page. You press a button and the script scrolls through the profile, collecting posts, images, video URLs, timestamps, and like counts, and stops automatically when it stops finding new posts. There are a few options you can set before starting. A scroll delay controls how many milliseconds the script waits between scrolls (1800 by default), with slower scrolls catching more text. A checkbox includes the profile's Replies tab in addition to the main feed. A separate Deep mode opens each post on its own page after the main scrape and pulls out the comment threads, keeping only the conversations where the profile owner actually replied. Deep mode is slow because it makes one extra request per post, so 200 posts takes about five to ten minutes. When the scrape finishes you can download the result as JSON, CSV, or Markdown. The JSON output includes the username, total counts, and an array of posts, each with its code, text, timestamp, like count, image URLs, a flag for whether it has video, and any captured conversations. The CSV is a flat spreadsheet-friendly version with one row per post. The Markdown export is meant to be readable, with headers and quoted conversations, and the README suggests dropping it into Notion or Obsidian. The README is honest about how the scraper works and what it cannot do. Threads only renders the posts currently in view, so the script scrolls in small steps and re-reads the DOM, merging information as posts re-appear. Deep mode parses hidden JSON inside the post pages. There is no login and no network interception, which is why it works alongside ad blockers, but it also means private accounts cannot be scraped, like counts may sometimes come through as zero if Threads does not expose them, and the script only runs on profile pages, not the home feed. It is written in plain JavaScript with Lucide icons, has no dependencies, and is released under the MIT license.
Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.