discovering-niche-podcasts
Discover niche/industry podcasts, podcast owners, interview episode sources, and guest/interviewee LinkedIn outreach lists using pasted reports, Perplexity, Exa, Tavily, and optional assisted Genspark
SKILL.mdpodcast-linkedin-intelligence-skills/skills/discovering-niche-podcasts/SKILL.md
GitHubdiscovering-niche-podcasts
Workflow
- Confirm the campaign niche, source list, or input file.
- Read
references/setup.mdif the repo is not configured. - Prompt for the minimum intake fields: niche, audience/ICP, optional seed terms, geography, must-include terms, and exclusions.
- Run the Apify SERP waterfall first with
scripts/run_podcast_intelligence_waterfall.py. - Use YouTube/RSS/transcript/OCR fallback scripts only for high-value shows whose guest names remain missing.
- Match both hosts/owners and interviewees to LinkedIn with evidence fields.
- Write descriptive outputs under
data/or.tmp/.
Scripts
scripts/discover_podcasts.pyscripts/ingest_research_report.pyscripts/match_linkedin_profiles.pyscripts/resolve_youtube_channels.pyscripts/scrape_youtube_podcast_guests.pyscripts/scrape_audio_podcast_guests.pyscripts/transcribe_audio_episodes.pyscripts/ocr_thumbnails.pyscripts/genspark_assisted.pyscripts/extract_youtube_urls_from_report.pyscripts/build_interviewee_linkedin_list.pyscripts/apify_serp.pyscripts/discover_podcast_interviews_serp.pyscripts/enrich_linkedin_via_apify_serp.pyscripts/run_podcast_intelligence_waterfall.py
References
references/research-report-workflows.mdfor turning pasted reports into podcast, episode, and LinkedIn enrichment outputs.
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.