LinkedIn Comment Replier

Use this template

Overview

This automation reads a LinkedIn post, finds real user comments using AI, and gets it ready to reply with your chosen message. It’s designed for long, messy pages and handles them reliably by breaking the content into smart chunks. Behind the scenes, it uses AI to tell genuine comments apart from LinkedIn’s UI (buttons, prompts, ads) and avoids replying to the original poster by mistake. It runs as a simple API you can call from your tools, and pairs nicely with a browser extension to act on LinkedIn through your browser.

Description

This automation extracts real comments from a LinkedIn post and prepares replies by combining the page content (captured through your browser) with AI technology into a clean, structured result. It first pulls the full page content for the post, then uses AI to identify each commenter’s name, profile link, and actual comment text while ignoring LinkedIn interface elements like reaction buttons and “People also viewed.” In parallel, it processes large posts in chunks so it stays fast and doesn’t hit model limits. After filtering out the original poster and deduplicating repeated profiles, it returns a ready-to-review list of comments and a summary so you can confidently send consistent replies at scale.

Key Features

Real comment extraction: Uses AI to find genuine user comments and ignore LinkedIn UI, reactions, ads, and prompts.

Smart chunking for long pages: Splits large pages into overlapping chunks to keep context and avoid missing replies.

Original poster filtering: Detects and skips the author’s own profile by URL and username to prevent accidental replies.

Deduplication by profile: Normalizes LinkedIn profile URLs and removes duplicates for clean, one-per-person results.

Parallel processing: Handles multiple chunks at once with safe concurrency limits for speed and reliability.

Robust token management: Conservatively estimates token usage and falls back to chunked processing if the content is huge.

Clear, structured output: Returns commenter name, profile URL, comment text, and a human-friendly summary you can review.

Configurable models and fallbacks: Uses Gemini 2.5 Flash via CodeWords Runtime with GPT‑4.1 as a quality fallback.

Instructions

1

Open the automation in CodeWords and connect your CodeWords account.

2

Install and sign in to the CodeWords browser extension so the automation can read the LinkedIn post.

3

Paste the LinkedIn post URL and type the reply message you want to use for each comment.

4

Hit run to let the automation scan the page, detect real comments, and prepare replies.

5

Review the results: check commenter names, profile links, and comment text, along with the summary.

6

Confirm and proceed to send replies (or adjust your message and run again) using the browser extension flow.

Use Cases

This automation is perfect for social media managers, founders, and customer success teams who want to acknowledge every comment on a LinkedIn post without spending hours doing it manually. It’s great for campaign launches, hiring announcements, product updates, and any post with lots of engagement where quick, consistent replies help you build relationships and keep the conversation going.

Frequently Asked Questions

Will this automation reply to the original poster by accident?

No. It detects the original poster by profile URL and username and skips them automatically.

How does it avoid replying to buttons or ads?

It uses AI to ignore LinkedIn interface elements like reactions, “Follow,” “Message,” sponsored content, and suggestion panels.

Can I preview comments before sending replies?

Yes. The automation returns a structured list of commenters with their profile links and comment text, so you can review first.

Does it work on very long posts with many comments?

Yes. It splits content into overlapping chunks and processes them in parallel, then deduplicates results.

What models does it use?

It runs Gemini 2.5 Flash through CodeWords Runtime and can fall back to GPT‑4.1 for quality and reliability.

How are duplicates handled?

Profile URLs are normalized and deduplicated so each commenter appears once, even if they show up in multiple chunks.

Can it handle different languages?

Yes. The AI technology can extract comments in multiple languages, and you can provide a reply message in the language you want.

Do I need to be logged in to LinkedIn?

Yes. You need access to view the post and reply. The browser extension helps the automation work through your logged-in session.

How long does it take?

Most posts process in under a couple of minutes. Very large threads may take longer as the automation safely batches content.

What do I need to provide to get started?

The LinkedIn post URL, the reply message you want to use, access via the browser extension, and your CodeWords API key.