Using Video Transcripts for AI Context
How to use your video's spoken transcript to help AI generate relevant, timed conversations.
What Is a Video Transcript?
A video transcript is the spoken words from your video — what the presenter says, discussions, explanations, and any verbal content. This is different from chat logs (the written messages from attendees).
When you provide a video transcript to FomoChat, it becomes the primary knowledge source for AI-generated conversations. The AI will:
- Reference topics only after they're mentioned in the video
- Generate questions and comments grounded in actual video content
- Create an authentic "watching together" experience
How It Works
- Paste your transcript in the Knowledge step (or Settings → Knowledge in edit mode)
- Set your video duration so timing scales correctly
- Generate your conversation — the AI uses the transcript to inform topics and timing
- Messages appear when relevant topics come up in the video
💡 Tip: The more accurate your transcript, the better the AI can generate relevant, well-timed conversations.
Where to Get Video Transcripts
Auto-Generated Captions
- YouTube: Open your video → Click "..." → "Open transcript" → Copy
- Zoom: Enable transcription in settings → Download from cloud recordings
- Vimeo: Check the "CC" section for auto-generated text
- Loom: Click the transcript icon in the sidebar
Transcription Services
- Otter.ai — Real-time and uploaded video transcription
- Rev.com — Human and AI transcription options
- Descript — Transcription with editing tools
- Trint — AI-powered transcription
Manual Transcription
For short videos, you can manually transcribe or use voice-to-text tools.
Adding Your Transcript
During Creation (Knowledge Step)
- Select Video/Webinar as your chat type
- Paste your transcript in the Video Transcript field
- Set your video duration in minutes
- Continue to generate participants and conversation
In Edit Mode (Settings → Knowledge)
- Open your FomoChat
- Go to Settings → Knowledge
- Paste or update your transcript
- Save changes
- Regenerate conversation if needed
Transcript Tips
Include Timestamps (Optional)
If your transcript has timestamps, keep them — they help the AI understand pacing:
[00:00] Welcome everyone to today's webinar...
[02:30] Let's dive into our first topic: pricing strategies...
[15:00] Now I'll show you a quick demo...
Keep It Accurate
- Fix obvious transcription errors
- Include speaker names if multiple presenters
- Keep filler words minimal but natural
Focus on Key Content
For very long videos, you can:
- Include the full transcript for best results
- Or focus on key sections you want the chat to reference
How Transcript Affects AI Generation
| Aspect | Behavior |
|---|---|
| Topic timing | Questions about pricing appear after pricing is discussed |
| Vocabulary | AI uses terms and phrases from the video |
| Relevance | Comments reflect what's being shown/said |
| Authenticity | Creates "I'm watching this live" feeling |
Video Transcript vs Chat Log Import
| Feature | Video Transcript | Chat Log Import |
|---|---|---|
| Source | Spoken words from video | Written chat from live event |
| Purpose | AI knowledge context | Verbatim message replay |
| Messages | AI-generated | Imported exactly |
| Best for | New/pre-recorded content | Replaying live events |
Related Articles
- Importing Webinar Chat Logs — Import actual chat messages instead
- Syncing Chat with Video — Time your conversation to video
- Setting Up Domain Knowledge — Other knowledge sources
Was this article helpful?
Your feedback helps us improve our documentation.