Podcast episode to social clips, show notes, and email in one workflow

Every week, you record a podcast. Every week, you spend eight hours turning that episode into something worth promoting: extracting clips, writing show notes, crafting email summaries, maybe even creating voiceovers for social media. You're not alone. Most podcast producers treat post-production as a separate job entirely, one that eats into their content strategy time. The problem isn't that these tasks are difficult. It's that they're repetitive, they're sequential, and they're manual. Your audience has already heard the episode. They're waiting for clips on TikTok, show notes in their inbox, and highlights in the newsletter. By the time you've finished editing the first clip by hand, you've already lost momentum. What if your podcast worked like this instead: upload the audio file, walk away, and return to find fully transcribed episodes, ten social clips ready to post, polished show notes with timestamps, and email copy that's already written. No toggling between five different windows. No copy-pasting timestamps. No "just one more manual review." This workflow does exactly that, using readily available tools connected through a single orchestration engine. For more on this, see Automated Podcast Production Workflow.

The Automated Workflow

The foundation of this system is n8n, an open-source workflow automation platform that handles orchestration without the monthly cost of Zapier or Make. You'll trigger the workflow via webhook when you upload a file to a cloud storage service, then chain together four specialist tools that each do one job well.

Step 1: Transcription and Show Notes

Start with Shownotes AI, which accepts audio files and returns structured transcription, timestamps, and summary data. When you upload your podcast episode to a designated folder in Google Drive or Dropbox, n8n catches the webhook and passes the file to Shownotes AI's REST API.

CopyPOST https://api.shownotesai.com/transcribe
Content-Type: application/json { "audio_url": "https://storage.example.com/podcast_ep_42.mp3", "language": "en", "output_format": "detailed", "include_summary": true
}

Shownotes AI processes the audio and returns a JSON response containing the full transcript, key moments with timestamps, and a paragraph-length summary. Store this response in your n8n workflow context; you'll use the transcript and summary multiple times downstream.

Step 2: Generating Clip Suggestions and Extraction

Feed the transcript into Claude Opus 4.6 via the Anthropic API. Claude's instruction-following is precise enough to identify natural clip boundaries, suggest compelling quotes, and even recommend which sections work best for different platforms.

CopyPOST https://api.anthropic.com/v1/messages
Content-Type: application/json { "model": "claude-opus-4.6", "max_tokens": 2000, "messages": [ { "role": "user", "content": "Here's a podcast transcript with timestamps. Identify 8-10 distinct clip segments (30 seconds to 2 minutes each) that work for short-form video. For each clip, provide: start time (in seconds), end time, platform recommendation (TikTok/Instagram/YouTube), and a one-sentence hook.\n\nTranscript:\n[FULL_TRANSCRIPT]" } ]
}

Claude returns structured suggestions. Parse the start and end times, then pass them to Clipwing's API, which handles the actual video cutting. If you already have a video version of your episode (many podcasters do), Clipwing accepts the source file and creates individual clips based on the timestamps you provide.

CopyPOST https://api.clipwing.com/create_clips
Content-Type: application/json { "source_video_url": "https://storage.example.com/podcast_ep_42.mp4", "clips": [ { "start_time": 145, "end_time": 205, "title": "Why remote work is broken", "platform": "tiktok" }, { "start_time": 420, "end_time": 520, "title": "The future of AI in hiring", "platform": "instagram" } ]
}

Clipwing stores the clips in your designated output folder and returns URLs. Pass these URLs and the clip titles to your social media scheduler (or store them in a spreadsheet for manual review before posting).

Step 3: Show Notes and Email Copy

Use the transcript and summary from Step 1 to generate two pieces of written content. HyperWrite's API accepts the transcript and returns polished show notes with proper formatting and embedded timestamps. Meanwhile, feed the same data to GPT-4o mini, which is fast and cost-effective for generating promotional email copy.

CopyPOST https://api.hyperwrite.com/generate_show_notes
Content-Type: application/json { "transcript": "[FULL_TRANSCRIPT]", "summary": "[SUMMARY_FROM_SHOWNOTES_AI]", "podcast_title": "Your Podcast Name", "episode_number": 42, "guest_name": "Guest Name (optional)"
}

For email generation, use GPT-4o mini to keep costs down:

CopyPOST https://api.openai.com/v1/chat/completions
Content-Type: application/json { "model": "gpt-4o-mini", "messages": [ { "role": "user", "content": "Write a 150-word email announcing this podcast episode to subscribers. Include a hook, one key takeaway, and a call-to-action to listen. Here's the episode summary:\n\n[SUMMARY]" } ]
}

Step 4: Optional Audio Voiceovers for Clips

For Instagram Reels or TikToks, some clips work better with a voiceover promoting the episode. ElevenLabs Turbo v2.5 generates natural-sounding speech from text. After Claude generates clip hooks, send those hooks to ElevenLabs to create brief voiceover MP3 files, then have n8n overlay them on the video clips using ffmpeg (called via a system command within n8n).

CopyPOST https://api.elevenlabs.io/v1/text-to-speech/21m00Tcm4TlvDq8ikWAM
Content-Type: application/json { "text": "Why remote work is broken. Listen to episode 42.", "voice_settings": { "stability": 0.75, "similarity_boost": 0.75 }
}

Step 5: Output and Storage

n8n writes all outputs to a structured folder in Google Drive: one subfolder for clips, one for show notes (as markdown), one for email copy, and one for voiceover files. Set up a final step that sends you a summary email with links to all generated content.

Copyjson
{ "episode": "42", "clips_generated": 9, "clips_folder": "https://drive.google.com/...", "show_notes_link": "https://drive.google.com/...", "email_draft": "Subject: Episode 42 , Why Remote Work Is Broken", "voiceovers_generated": 9
}

Set this workflow to run on a schedule (e.g. whenever a new audio file appears in your input folder) or trigger it manually. The entire process takes 5-10 minutes for a typical 45-minute episode, with zero manual handoff between tools.

The Manual Alternative

If you prefer more control over which clips make the cut or want to add custom branding to each video, you can run steps 1 and 2 automatically, then review Claude's clip suggestions before instructing n8n to proceed with extraction and voiceover generation. This takes 20-30 minutes instead of 5 but gives you editorial control without full manual re-creation. Alternatively, skip Clipwing entirely and download the timestamps, then use a local tool like FFmpeg directly. This works well if you already have a video editing workflow in place.

Pro Tips

Error handling and retries.

Shownotes AI occasionally times out on very long files.

Configure n8n to retry failed transcription requests up to three times with exponential backoff (5 seconds, 10 seconds, 20 seconds). Set a webhook timeout of 60 seconds minimum.

Rate limits and costs.

Anthropic charges per token, and Claude Opus is the most expensive. Use Claude Sonnet 4.6 instead if you're running this workflow multiple times per week; the clip suggestions are nearly as good and you'll cut costs by 60 percent. ElevenLabs charges per character of generated speech; batch voiceover generation at night or use their Turbo v2.5 plan with higher monthly allowance if you're processing more than two episodes per week.

Testing the chain.

Before running this on a full episode, test with a 10-minute segment first. This catches API key issues, storage permission problems, and data format mismatches before you waste time on a 90-minute file.

Storing API responses.

Save each intermediate output (transcript, clip suggestions, show notes) as a JSON file in your n8n project. This makes debugging easier and lets you re-run downstream steps without re-processing earlier ones if something fails.

Platform-specific formatting.

Clipwing outputs video clips, but platform specs vary: TikTok prefers vertical video (9:16), Instagram Reels accept both vertical and square (1:1), YouTube Shorts work at multiple aspect ratios. Configure Clipwing or ffmpeg to format each clip appropriately before storing it.

Cost Breakdown