README.md

metadata

license: cc-by-nd-4.0
task_categories:
  - text-generation
  - text-classification
  - feature-extraction
  - question-answering
tags:
  - finance
  - nlp
  - transcripts
  - ceo
  - executives
  - sentiment-analysis
  - hedge-fund
  - alternative-data
  - earnings-calls
  - podcasts
  - interviews
  - market-signals
  - quant
  - fed
  - macro
size_categories:
  - 100K<n<1M

🎙️ CEO Transcripts — Verified Executive Interviews

The World's Largest Database of Verified C-Suite Transcripts

20,000+ Executives · 100,000+ Transcripts · 400,000+ Quotes · S&P 500 + NASDAQ + Global Leaders

---

🔥 What's In This Sample?

This is a free evaluation sample from CEOInterviews.ai featuring 9 of the most market-moving voices in finance, tech, and policy.

Executive	Role	Why They Matter
Jensen Huang	CEO, NVIDIA	Every AI demand signal moves $3T in market cap
Jerome Powell	Chair, Federal Reserve	His words move trillions—quants build NLP models on Fed language
Elon Musk	CEO, Tesla	Highest search volume globally, 3-hour podcast appearances
Warren Buffett	CEO, Berkshire Hathaway	Most famous investor alive, legendary annual meeting Q&As
Donald Trump	45th & 47th President	Most searched political figure globally, 3-hour Rogan appearance, tariff policy signals move markets
Mark Zuckerberg	CEO, Meta	Podcast king—Rogan, Fridman, unscripted strategy reveals
Jamie Dimon	CEO, JPMorgan	Banking bellwether, "hurricane" recession calls move markets
Cathie Wood	CEO, ARK Invest	Retail trader icon, daily commentary on disruptive innovation
Ray Dalio	Founder, Bridgewater	Hedge fund legend, macro frameworks every quant knows

Sample Date Range: 3 months (August — Nov 2025) Full Dataset: 2007-Present (updated daily)

---

💡 The Problem We Solve

Alpha lives in unguarded moments.

The media playbook for executives has fundamentally shifted:

Old Playbook	New Reality
Scripted earnings calls	3-hour Joe Rogan podcasts
PR-approved press releases	Unscripted Lex Fridman interviews
Quarterly investor days	Off-the-cuff conference Q&As

The signal is there. But it's buried in 10,000+ hours of fragmented audio.

CEOInterviews.ai transforms this chaos into structured, queryable, backtestable data.

---

🎯 Example: Powell vs Dimon on Recession Risk

Research Question: "What did Jerome Powell and Jamie Dimon say about recession risk in 2022?"

Using Our API:

import requests

API = "https://ceointerviews.ai/api"
headers = {"X-API-Key": "your_api_key"}

powell_quotes = requests.get(f"{API}/quotes/", params={"entity_id": 15847, "keyword": "recession"}, headers=headers).json()
dimon_quotes = requests.get(f"{API}/quotes/", params={"entity_id": 12903, "keyword": "recession"}, headers=headers).json()

for q in powell_quotes["results"][:3]:
    print(f"[Powell {q['source_created_at'][:10]}] {q['text'][:150]}...")

for q in dimon_quotes["results"][:3]:
    print(f"[Dimon {q['source_created_at'][:10]}] {q['text'][:150]}...")

Sample Output:

[Powell 2022-06-15] "We're not trying to induce a recession now, let's be clear about that. 
We're trying to achieve 2% inflation..."

[Dimon 2022-06-01] "You know, I said there's storm clouds but I'm going to change it... 
it's a hurricane. Right now, it's kind of sunny, things are doing fine..."

[Powell 2022-09-21] "We have got to get inflation behind us. I wish there were a painless 
way to do that. There isn't..."

[Dimon 2022-09-26] "This is serious stuff... it's a different environment than we've ever 
seen before... the Fed has to meet this now..."

This is the alpha. Dimon called the "hurricane" 2 weeks before Powell acknowledged pain was coming.

---

📊 Dataset Schema

Each row is a notable quote with full context about the executive and source video.

Field	Type	Description
quote_text	string	The quote itself
entity_simple_name	string	Normalized name (e.g., "Jensen Huang")
entity_title	string	Job title
entity_profile_pic_url	string	Profile image URL
post_video_thumbnail_url	string	Video thumbnail URL
post_appearance_date	date	When executive actually spoke ⭐
company_logo_url	string	Company logo URL
entity_institution	string	Company/organization
post_title	string	Video/podcast title
post_source_url	string	Original YouTube URL
quote_is_notable	bool	AI-flagged as significant
quote_is_controversial	bool	AI-flagged as controversial
quote_is_financial_policy	bool	Market/policy relevant
quote_topics	list	Extracted topics (e.g., ["tariffs", "trade"])
quote_entities	list	People mentioned
quote_companies	list	Companies mentioned
post_description	string	Video description
post_source_created_at	datetime	When video was published
quote_timestamp_from_tx	string	Timestamp in video (e.g., 00:01:48,000)
quote_created_at	datetime	When we extracted this quote
entity_id	int	Executive ID
entity_name	string	Full name
entity_is_politician	bool	Is a political figure
entity_is_company_leader	bool	Is a corporate executive
entity_is_top_influencer	bool	High-profile individual
company_id	int	Company ID
company_ticker	string	Stock ticker (e.g., "NVDA")
post_id	int	Source video/transcript ID
post_appearance_lag_days	int	Days between appearance and publish
post_thumbnail_quality_review_score	float	AI thumbnail verification (0-1)
post_transcript_quality_review_score	float	AI transcript verification (0-1)
post_transcript_length_chars	int	Transcript length in characters
post_has_transcript	bool	Has full transcript available
transcript_demo_only_request_licensed_access	list	Transcript preview (first 20 lines)
quote_id	int	Unique quote identifier

---

⭐ What Makes CEOInterviews Different?

1. Appearance Date Detection

Most datasets only have publish_date. But a podcast uploaded today might contain an interview from 6 months ago.

We detect when the executive actually spoke.

# Example: Interview recorded in January, published in March
{
    "publish_date": "2022-03-15",      # When YouTube uploaded
    "appearance_date": "2022-01-20",   # When Buffett actually spoke ⭐
}

This is critical for backtesting. You need to know when the market could have known, not when the video appeared.

2. AI + Human Verification

Every transcript passes through:

• 🤖 Thumbnail Analysis: Is the executive actually in the video?
• 🤖 Transcript Verification: Is this their voice, not dubbed/AI-generated?
• 🤖 Quality Scoring: Is the transcript complete and accurate?

No deepfakes. No dubbing. No secondhand reporting.

3. Structured Quote Extraction

We don't just give you transcripts—we extract the market-moving moments:

{
    "text": "We're seeing something we haven't seen in 40 years...",
    "is_notable": True,
    "is_financial_policy": True,
    "topics": ["inflation", "monetary_policy", "fed"],
    "timestamp_in_video": "00:23:45"
}

---

📈 Use Cases

Use Case	How This Dataset Helps
Sentiment Alpha	Backtest NLP signals on Fed language, CEO confidence
Event Studies	Measure market reaction to specific statements
ESG Tracking	Monitor executive commitments on climate, DEI
Competitive Intel	What is Jensen saying about AMD? What is Zuckerberg saying about Apple?
Macro Signals	Track Dalio's "changing world order" thesis evolution
LLM Fine-tuning	Train models on authentic executive communication
Media Analysis	How do CEOs communicate differently on Rogan vs CNBC?

---

📊 Sample vs Full Dataset

Metric	This Sample	Full Dataset
Executives	9	20,000+
Transcripts	~1400	100,000+
Quotes	~7,000	400,000+
Date Range	2025	2007-Present
Updates	Static	Daily
API Access	❌	✅ 1,000 req/min
Custom Entities	❌	✅ On request

---

🔐 Get Full Access

This sample is <1% of our full dataset.

Full Dataset Includes:

• ✅ Every S&P 500 and NASDAQ CEO
• ✅ Global political leaders (Presidents, Prime Ministers, Fed chairs)
• ✅ Top AI founders (Altman, Hassabis, etc.)
• ✅ Legendary investors (Buffett, Dalio, Ackman, Fink)
• ✅ Daily updates
• ✅ RESTful API with full pagination
• ✅ CSV/JSON export for model training
• ✅ White-glove enterprise support

Pricing

Tier	Price	Best For
Researcher	$499/mo	Academic research, small teams
Professional	Custom	Hedge funds, trading desks
Enterprise	Custom	Full API access, custom coverage

Contact

📧 Email: lucas@ceointerviews.ai
🌐 Website: ceointerviews.ai
📚 API Docs: ceointerviews.ai/api_docs

---

📜 License

CC-BY-NC-ND-4.0 (Creative Commons Attribution-NonCommercial-NoDerivatives 4.0)

What This Means:

✅ Allowed	❌ Not Allowed
Download and evaluate	Commercial use
Academic research	Redistribute modified versions
Personal projects	Resell or sublicense
Cite in papers	Create derivative datasets

For commercial use, contact lucas@ceointerviews.ai

---

🛠️ Built By

Lucas Ou-Yang
Former Engineering Manager and Staff ML Engineer @Coinbase, @Tiktok, @Meta superintelligence labs

Building institutional-grade datasets for quantitative research.

---

📣 Citation

If you use this dataset in research, please cite:

@misc{ceointerviews2024,
  author = {Ou-Yang, Lucas},
  title = {CEOInterviews.ai: Verified Executive Transcript Dataset},
  year = {2024},
  publisher = {HuggingFace},
  url = {https://huggingface.co/datasets/codelucas/ceo-transcripts-verified-sample}
}

---

⭐ Star this dataset if you find it useful!

Questions? Reach out at lucas@ceointerviews.ai