Reach — LinkedIn & YouTube Visibility Tool by Prateek SahniReach — LinkedIn & YouTube Visibility Tool by Prateek Sahni

Reach — LinkedIn & YouTube Visibility Tool

Portfolio project
Prateek Sahni

The Problem

Commenting early on popular posts is a good way to be seen, but finding the right posts every day and writing something useful each time takes hours, and automated commenting is spam and against LinkedIn's terms.

What I Built

A command-line tool in Go that searches six topics for LinkedIn posts and YouTube videos from the last day, keeps only near-viral ones (20+ comments, and 10,000+ views on YouTube), and ranks them by how fast they are climbing, audience size and how visible a new comment would be.
Scoring formula: Velocity (engagement / age), reach (log-scaled follower count) and visibility (fewer existing comments = more visible). Posts older than 24 hours are skipped.
LLM drafting: An LLM chain on NVIDIA's API drafts each comment from the real projects on my portfolio site. Hard rules check every draft: 200 characters or fewer with one allowed link, no generic openers, hype or hashtags, and every number must appear in the post or the project facts. Drafts that fail are sent back up to three times.
Human approval: Drafts go to a Google Sheet for review. An assist command opens each approved post with the comment copied, so I post it myself. Nothing is ever posted automatically.
YouTube auto-post: YouTube comments can be submitted through the official API with my own sign-in, respecting moderation. LinkedIn stays manual (its comment API is partner-only).

The Outcome

A daily shortlist of the posts most worth commenting on, with fact-checked drafts ready to approve. Every comment links back to real work. Open source.

0 likes · 0 views · Nov 15, 2026

Go tool that finds fast-rising LinkedIn and YouTube posts, drafts fact-checked comments and queues them for human approval.

Timeline

Jan 1, 2026 - Ongoing

Tags