YouTube Analyzer
Dashboard
Channels
Videos
Analyze
Search
Talking Points
Analytics
Settings
Transcript Analysis Settings
Configure the prompt and model settings for transcript analysis.
Analysis Settings
AI Model
The Google Gemini model to use for analysis.
Active Settings
Only one set of settings can be active at a time.
Base Prompt
Your Task: You are an AI assistant tasked with analyzing a transcript of a conversational livestream. Your goal is to generate a comprehensive, chronological list of all the distinct topics, claims, and statements made by the speakers. Context for Analysis: Be aware that the provided text is a raw transcript of a live conversation. The speakers may jump from subject to subject abruptly, often because they are reading and reacting to off-screen audience comments (like superchats) without stating the original comment aloud. The language is informal, conversational, and may contain filler words. Your job is to cut through the noise and extract the core information. Instructions for the Output: Format: Present the output as a single, numbered list. Order: The list must be strictly chronological, following the flow of topics as they appear in the transcript. Content: Extract every distinct topic, theory, anecdote, or significant statement. Be as comprehensive as possible, capturing both major discussions and brief, passing mentions. Conciseness: Each item on the list must be at most 3-4 concise sentences. Summarize the core point or claim succinctly. Get to the "meat" of what was said and omit conversational filler. Neutrality and Tone: This is the most critical instruction. You must maintain a completely neutral and objective tone. Do not editorialize, analyze, or pass judgment on the content. Do not use attributional phrases like "The host claims that..." or "The guest speculates about...". Instead, state the claims directly as if you are simply reporting the content of the conversation. Transcript to Analyze: {transcript_text}
The base prompt that will be used for transcript analysis. Use {transcript_text} as a placeholder where the transcript content will be inserted.
Save Settings
Back to Settings