scraping-audio-podcast-guests

Workflow

1. Confirm the campaign niche, source list, or input file.
2. Read references/setup.md if the repo is not configured.
3. Prefer the Apify SERP waterfall for discovery; it can find Apple/Spotify/RSS-only episode pages before transcript spending.
4. Run title-first extraction with scripts/build_interviewee_linkedin_list.py or scripts/scrape_audio_podcast_guests.py.
5. Use transcript extraction only for important shows where titles/descriptions omit guest names.
6. Write descriptive outputs under data/ or .tmp/.
7. Preserve evidence and candidate fields for review.

Scripts

• scripts/discover_podcasts.py
• scripts/ingest_research_report.py
• scripts/match_linkedin_profiles.py
• scripts/resolve_youtube_channels.py
• scripts/scrape_youtube_podcast_guests.py
• scripts/scrape_audio_podcast_guests.py
• scripts/transcribe_audio_episodes.py
• scripts/ocr_thumbnails.py
• scripts/genspark_assisted.py
• scripts/build_interviewee_linkedin_list.py
• scripts/run_podcast_intelligence_waterfall.py

Guardrails

• Never hardcode API keys.
• Never accept low-confidence LinkedIn matches automatically.
• Do not scrape LinkedIn directly unless the user provides a compliant workflow.
• Keep raw private research reports out of git unless sanitized.