How to Access Your Stored Transcript Library Through AI

Everything you process through TokScript - whether on the website, through the Chrome extension, or through the AI connector - gets stored in your TokScript library. The MCP connector gives you full access to search, browse, and retrieve that stored data from within Claude or ChatGPT.

Why This Matters

If you transcribed 50 videos last week, you don't need to re-scan them. The data is already in your library, and the AI can pull it instantly. This opens up historical analysis that isn't possible with live-only scanning: trend tracking over weeks or months, comparing your content performance across time periods, and querying your entire video archive with natural language.

Getting Library Stats

Ask for an overview of your entire library to see total videos stored, breakdown by platform, folder counts, your most-scanned authors, date range of your content, and recent activity.

Example prompts: "How many videos do I have in my TokScript library?" or "Show me my library stats" or "Give me an overview of my stored transcripts"

Searching and Browsing Your Library

You can search across titles, author names, transcript text, tags, and descriptions. Filter by platform (TikTok, Instagram, YouTube), by folder, or by bookmarked items. Sort results by newest, oldest, title alphabetically, or most viewed.

Example prompts: "Search my library for videos about cooking" or "Show me all my Instagram transcripts" or "What's in my 'Competitors' folder?" or "Show me my most-viewed stored videos" or "Find all bookmarked TikToks in my library"

Retrieving Full Data for a Stored Video

When you want the complete stored data for a specific video - including the full transcript, all engagement stats, audio information, and everything else - you can pull it by URL or by the document ID from your library.

Example prompts: "Get the full stored data for this video: [URL]" or "Show me the complete transcript I have saved for [video title]"

The AI checks your library first before making any live requests, so retrieving stored data is fast - it's a direct database read, not a fresh scan.