Pull 20 YouTube Videos Into a Searchable Knowledge Base

Step-by-step tutorial: connect your agent to Felo OpenAPI, extract YouTube transcripts, analyze key topics, and build a searchable team knowledge base in LiveDocs.

Your team has bookmarked dozens of technical videos — conference talks, tutorials, product demos. But they're locked in video format. No one has time to watch them all.

Here's how to extract the knowledge and make it searchable.

3 Minutes to Connect Felo OpenAPI

Step 1: Go to openapi.felo.ai and create a free API key.

Step 2: Copy and paste this into your Codex, Claude Code, or OpenClaw:

Connect to Felo OpenAPI:
- Models: Claude Opus 4.8, Claude Sonnet 4.6, Claude Haiku 4.5, GPT-5.6 Sol, GPT-5.6 Terra, Grok 4.5
- Search: POST /v2/chat (AI Search), POST /v2/web/extract (Web Fetch)
- Video: GET /v2/youtube/subtitling (YouTube Subtitles)
- Output: POST /v2/ppts (PPT), POST /v2/mindmap, POST /v2/landing_page, Image Generation
- Memory: LiveDocs (semantic retrieval, file upload, URL resources, tasks, records)
Base URL: https://openapi.felo.ai
API Key: [YOUR_API_KEY]

When I give you YouTube URLs:
1. Extract transcripts via YouTube Subtitles API
2. Analyze key topics, technical points, and takeaways
3. Save structured summaries to LiveDocs with video metadata
4. Generate a knowledge map via Mindmap API

Step 3: Connection complete. Your agent can now extract video content and build knowledge bases.

The User Story

Your engineering team shared 10 conference talk videos from a recent AI conference. You want to:

• Extract the full content from each video
• Identify key technical points and actionable insights
• Organize everything into a searchable knowledge base
• Create a visual map of how topics connect

No one should have to watch 8 hours of video to find the one technique they need.

You Only Need to Say This

Extract transcripts from these YouTube videos, analyze the key topics and technical insights, and build a searchable knowledge base in LiveDocs. Also generate a knowledge map mindmap.

Video URLs:
1. https://youtube.com/watch?v=abc123 — "The Future of AI Agents"
2. https://youtube.com/watch?v=def456 — "Building Production RAG Systems"
3. https://youtube.com/watch?v=ghi789 — "LLM Evaluation Benchmarks"
[... more URLs ...]

What Codex Does Automatically

Step 1: Extract Transcripts

For each video, Codex calls:

GET /v2/youtube/subtitling?url=https://youtube.com/watch?v=abc123

Returns: Full transcript with timestamps.

Codex processes each URL one by one, collecting all transcripts.

Step 2: Analyze Each Video

For each transcript, Codex extracts:

• Main topic: What is this video about?
• Key technical points: What specific techniques, tools, or frameworks are discussed?
• Actionable insights: What can the team implement based on this talk?
• Quotes: Notable statements worth preserving
• References: Tools, papers, or resources mentioned

Step 3: Save to LiveDocs

Codex saves each video as a structured entry:

Title: The Future of AI Agents
Source: https://youtube.com/watch?v=abc123
Speaker: [Name]
Duration: 45 minutes
Main Topic: AI agent architecture and capabilities

Key Technical Points:
- Multi-agent coordination patterns
- Tool calling protocol design
- Memory management in long-running agents

Actionable Insights:
- Implement shared memory pool for agent teams
- Use structured tool definitions for better agent coordination

Notable Quotes:
- "The best agent is one that knows when to delegate"

Tags: [AI agents, multi-agent, architecture, conference-2026]

Step 4: Generate the Knowledge Map

POST /v2/mindmap
{
  "topic": "AI Conference Knowledge Map",
  "content": [all analyzed video topics and relationships]
}

Returns: A visual mindmap showing how topics from different videos connect — for example, "agent memory" discussed in 3 different talks, or "evaluation methods" covered from multiple angles.

Step 5: Create a Summary Presentation

POST /v2/ppts
{
  "topic": "AI Conference — Key Takeaways",
  "content": {
    "overview": [number of videos, total hours, main themes],
    "top_insights": [most important findings across all videos],
    "action_items": [things the team should implement],
    "resource_links": [papers, tools, and frameworks mentioned]
  },
  "theme": "conference-summary"
}

Returns: A summary presentation with the best insights from all videos.

The Result

Your team now has:

• Searchable Knowledge Base: Every video's content indexed in LiveDocs. Anyone can search for a topic and find the relevant video, exact timestamp, and key quotes.
• Knowledge Map: A visual mindmap showing how topics connect across videos.
• Summary Presentation: The best insights from all videos in a shareable deck.
• Action Items: Specific techniques the team should explore or implement.

Instead of 8 hours of video watching, your team gets 10 minutes of curated knowledge.

The Complete Prompt Template (Copy & Reuse)

Extract transcripts from these YouTube videos and build a knowledge base:

[Video URL 1] — [Title/description]
[Video URL 2] — [Title/description]
[Video URL 3] — [Title/description]

For each video:
1. Extract the transcript
2. Analyze key topics, technical points, and actionable insights
3. Save a structured summary to LiveDocs with tags

Then:
4. Generate a knowledge map mindmap showing topic connections
5. Create a summary presentation with the top insights

Replace the URLs and titles for any video collection.

Advanced: Continuous Learning

Set up a recurring prompt:

Check our LiveDocs knowledge base for videos about [TOPIC]. Search YouTube for any new videos on this topic from the past month. Extract transcripts for new videos, add them to the knowledge base, and flag any new insights that weren't in the previous collection.

Getting Started

1. Create your free API key
2. Paste the initialization prompt into your agent
3. Try the video knowledge base prompt above

Unlock the expertise hidden in video. Let your agent watch, extract, and organize.

Get Your Free API Key →