___  ___  ___/ / _/ /__ _    __
   / _ \/ _ \/ _  / _/ / _ \ |/|/ /
  / .__/\___/\_,_/_//_/\___/__,__/
 /_/

Your podcast intelligence and guest pitching tool for music PR. Extract guests, episode angles, host matches, and insights from Apple Podcasts and Spotify RSS feeds.

Part of the Total Audio Promo suite (totalaudiopromo.com).

---

Workspace Packages

• podflow CLI: CLI digest tool & native MCP server (src/cli.ts, npm package podflow).
• @podflow/web: Next.js web application (apps/web) featuring the interactive pitching simulator, host match scoring cards, and PR roster dashboard.
• @totalaudiopromo/ui: Total Audio Promo shared design system (packages/ui) with Studio Purple (#A855F7) theme tokens, Tailwind presets, and React components.

---

Web App Quick Start

pnpm install
pnpm --filter @podflow/web dev    # starts Next.js dev server on port 3010

---

Quick Start (CLI)

npx podflow init                          # create config
npx podflow digest --dry-run              # preview episodes
npx podflow digest --max-episodes 10      # process 10 episodes

Or install globally:

npm install -g podflow
podflow digest

Commands

Command	Description
podflow init	Create config at ~/.podflow/
podflow digest	Process new episodes (default)
podflow digest --dry-run	Preview without API calls
podflow digest --backfill	Process all episodes
podflow subs	List podcast subscriptions
podflow stats	Cache statistics
podflow schedule	Weekly automatic digests + notification (macOS)
podflow mcp	Start the MCP server (stdio) — ears for agents

Set it and forget it

podflow init now walks you through setup interactively (provider, key, interests) — no JSON editing. Then:

podflow schedule            # weekly, Mondays 08:00
podflow schedule --daily --time 07:30
podflow schedule --status   # check it
podflow schedule --off      # remove it

Each scheduled run processes new episodes and ends with a macOS notification, so the digest reaches you instead of waiting in a file. Your API key is stored in ~/.podflow/config.json (permissions 0600) so scheduled runs work without shell environment variables.

Agents can use it too (MCP)

podflow ships an MCP server, so AI agents can query your podcast intelligence directly — what was said, by whom, and why it matters.

{
  "mcpServers": {
    "podflow": {
      "command": "npx",
      "args": ["-y", "podflow", "mcp"]
    }
  }
}

Six tools:

Tool	What an agent gets
search_episodes	Keyword search across titles, guests, ideas, and notes
get_guest	Cross-episode guest profile: role, company, appearances, why follow
get_ideas	Extracted ideas filtered by topic and relevance
list_follow_worthy	Guests flagged worth following, with context
get_stats	Cache totals and cost
ingest_feed	Point it at any podcast RSS feed to extract new episodes on demand

Intelligence is episode-level (descriptions plus Apple transcript snippets where available). Uses your existing ~/.podflow config and provider key; ingest_feed costs pennies per episode, same as digest.

---

Why podflow?

• Local-first & Hosted. Reads Apple Podcasts SQLite database locally or runs via the hosted Podflow Web UI.
• Your AI key, your choice. Supports Anthropic, OpenAI, Google, and Ollama (local, free).
• Incremental. Only processes new episodes. Run it weekly, costs pennies.
• Configurable interests. Scores everything against what matters to you.

How It Works

1. Reads the Apple Podcasts SQLite database & RSS feeds
2. Prioritises episodes by podcast tier (you configure which podcasts matter)
3. Batches episodes and sends to your chosen AI provider
4. Extracts guests, ideas, people mentioned, and host match relevance scores
5. Caches results incrementally (interrupted runs resume)
6. Generates a markdown digest & powers the Podflow web pitching dashboard

Configuration

After podflow init, edit ~/.podflow/config.json:

{
  "about": "A founder building developer tools",
  "interests": [
    {
      "name": "Customer Acquisition",
      "keywords": ["pricing", "growth", "retention", "PLG"],
      "why": "Growing from 0 to first paying customers"
    },
    {
      "name": "AI & Agents",
      "keywords": ["AI", "LLM", "agent", "automation"],
      "why": "Building AI-powered features"
    }
  ],
  "provider": "anthropic",
  "model": "claude-haiku-4-5-20251001",
  "outputPath": "./podflow-digest.md"
}

Podcast Tiers

Edit ~/.podflow/podcasts.json to prioritise specific podcasts:

{
  "podcasts": {
    "My Favourite Podcast": { "tier": 1, "extractGuests": true, "extractIdeas": true },
    "Entertainment Only": { "tier": 4, "extractGuests": false, "extractIdeas": false }
  },
  "defaults": { "tier": 3, "extractGuests": true, "extractIdeas": true }
}

Tiers 1-3 are processed by default. Tier 4 is skipped (use --include-all to override).

Global Flags

--dry-run               Preview without API calls
--backfill              Process all episodes
--max-episodes <n>      Limit episodes (default: 50)
--tier <n>              Only process specific tier (1-4)
--include-all           Include tier 4
--provider <name>       anthropic, openai, google, ollama
--model <name>          Override model name
--output <path>         Output file path
--recent <days>         Days of history (default: 365)
--verbose               Show episode details
-q, --quiet             Suppress output except errors

AI Providers

Provider	Env Var	Cost / 1,000 eps	Best For
Anthropic (Haiku)	ANTHROPIC_API_KEY	~$2	Best quality/cost ratio
OpenAI (GPT-4o-mini)	OPENAI_API_KEY	~$0.50	Cheapest cloud option
Google (Gemini Flash)	GOOGLE_GENERATIVE_AI_API_KEY	~$0.25	Budget option
Ollama (local)	None	Free	Privacy, no API key

podflow digest --provider openai --model gpt-4o-mini
podflow digest --provider ollama --model llama3.2

Output

Generates a markdown digest with:

• People to Follow -- guests flagged as worth following, with roles, companies, and social links
• Ideas by Topic -- actionable insights grouped by your configured interest areas
• High Relevance -- episodes scored 7+ for relevance to your work
• Recent Insights -- last 30 days of extracted intelligence
• Guest Index -- every guest across all episodes, cross-referenced

Cost

Scenario	Episodes	Cost (Haiku)
Weekly run	~20	~$0.03
Monthly	~80	~$0.12
Full backfill	1,000	~$2.00

Requirements

• macOS with Apple Podcasts (Sync Library enabled)
• Node.js 20+
• An API key for your chosen provider (or Ollama for local)

Contributing

See CONTRIBUTING.md for dev setup and guidelines.

Changelog

See CHANGELOG.md for release history.

Part of Total Audio

Tools I build for music PR, by Chris Schofield. Part of Total Audio Promo.

Project	Description
TAP	Campaign management for music PR agencies
totalaud.io	Release planning for emerging artists
SpotCheck	Spotify playlist validation
Newsjack	Music industry newsjacking
Podflow	Podcast intelligence for music PR
Sink	Contact data hygiene CLI

Questions? Reach me on X/@chrisschouk or info@totalaudiopromo.com.

Licence

MIT