# Claude Code Web Search Only Reads Titles, Not Full Content

> A trending GitHub skill adds deeper search to Claude Code, pulling full Reddit, X, and YouTube transcripts and comments instead of just titles and metadata.

Published: 2026-07-30
URL: https://daniliants.com/insights/claude-code-web-search-only-reads-titles-not-full-content/
Tags: claude-code, ai-search, skills, web-scraping

---

## Summary

A TikTok promotes a trending GitHub repo called "Last 30 Days" that adds an alternative search capability to Claude Code, pulling full transcripts and comments from Reddit, Twitter, and YouTube instead of just titles or metadata. The creator claims this produces better sentiment and research insight than Claude Code's built-in web search, but defers the actual explanation to a separate YouTube video.

## Key Insight

- The core claim: Claude Code's default web search only surfaces thread titles and metadata, not full transcripts or comment threads, so sentiment analysis and deep research are shallow out of the box.
- The named tool ("Last 30 Days") is pitched as a skill or plugin that widens the source set to Reddit, Twitter (X), and YouTube, pulling full content rather than summaries.
- No concrete mechanism, install steps, pricing, or benchmark numbers are given in the clip itself. It functions as a teaser for a longer video, which is a known low-substance content pattern.
- Worth noting for context: this maps onto a real known limitation, since built-in web search operates at the metadata and title level. That is why dedicated research skills or subagents that fetch full page content and comments tend to outperform it.