Healthcare clinic patient education video creation

Patient education videos are one of the most effective ways to improve health outcomes, yet most clinics produce them manually or don't produce them at all. A typical workflow involves scripting, recording, editing, adding voiceovers, and publishing. Each step requires specialist software and often multiple people. The result: clinics either skip video education entirely or spend months producing a handful of videos.

This Alchemy workflow removes that friction entirely. You provide a medical topic, a script, or even rough notes, and the system generates a complete patient education video with professional narration and matching visuals. No video editing software required. No manual coordination between teams. The entire process runs automatically from trigger to published output.

We'll show you how to wire together ElevenLabs for natural-sounding voiceovers, Hour One for AI-generated presenter videos, and Pika AI for dynamic visual sequences. Using Zapier or n8n as your orchestrator, you'll create a workflow that transforms a Google Sheet entry into a finished video in less than an hour.

The Automated Workflow

Which Orchestration Tool to Use

For this workflow, we recommend n8n if your clinic has any technical capacity; Zapier if you want the simplest setup with minimal maintenance. Both will work equally well. n8n is free to self-host and offers unlimited executions, making it cheaper long-term. Zapier requires less setup but counts every execution against your plan limit.

We'll provide examples for both below.

The Complete Data Flow

Your workflow follows this sequence:

1. Clinic staff adds a row to a Google Sheet with: topic, script or outline, target audience, and any visual preferences.

2. The orchestrator detects the new row and passes the script to ElevenLabs.

3. ElevenLabs generates an MP3 voiceover with a professional healthcare voice.

4. The system sends the script and voiceover metadata to Hour One.

5. Hour One generates a presenter video with the script delivered by an AI avatar.

6. The system extracts key phrases from the script and sends them to Pika AI.

7. Pika AI generates supporting visuals (animations, anatomical diagrams, or scenes).

8. The orchestrator combines the Hour One video with Pika AI visuals and embeds the voiceover.

9. The finished video uploads to YouTube, Vimeo, or your clinic's video server.

10. The orchestrator updates the Google Sheet with the video link and completion status.

Setting Up in n8n (Recommended)

Start by creating a new workflow. You'll need API keys from all three tools.

ElevenLabs provides a free tier with 10,000 characters monthly. Hour One offers a free tier with limited avatars. Pika AI gives $30 monthly free credits.

Connect your Google Sheet trigger first. Use the "Google Sheets" node and configure it to watch a specific worksheet for new rows.

Copy
{
  "resource": "spreadsheet",
  "operation": "appendRow",
  "spreadsheetId": "YOUR_SHEET_ID",
  "range": "Sheet1",
  "values": [
    "{{ $json.topic }}",
    "{{ $json.script }}",
    "{{ $json.audience }}",
    "{{ $json.videoUrl }}"
  ]
}

Next, add the ElevenLabs node. ElevenLabs has a straightforward REST API for text-to-speech. You'll POST your script text and receive an audio file URL.

Copy
POST https://api.elevenlabs.io/v1/text-to-speech/{{ voiceId }}
Headers:
  xi-api-key: YOUR_ELEVENLABS_KEY
  Content-Type: application/json

Body:
{
  "text": "{{ $json.script }}",
  "model_id": "eleven_monolingual_v1",
  "voice_settings": {
    "stability": 0.5,
    "similarity_boost": 0.75
  }
}

Choose a professional healthcare voice. ElevenLabs voice IDs like "Rachel" (ID: 21m00Tcm4TlvDq8ikWAM) work well for patient education. Store the returned audio URL in a workflow variable....... For more on this, see Healthcare patient education video creation from clinical....

The Hour One API accepts a script and generates a presenter video. This is your visual anchor, the person delivering the information. Create a new HTTP Request node in n8n targeting Hour One's API.

Copy
POST https://api.hourone.ai/videos
Headers:
  Authorization: Bearer YOUR_HOURONE_TOKEN
  Content-Type: application/json

Body:
{
  "script": "{{ $json.script }}",
  "avatar_id": "anna_la_uk",
  "voice_id": "{{ voiceFromElevenLabs }}",
  "quality": "high",
  "output_format": "mp4"
}

Hour One's response includes a video_id and polling URL. You'll need to poll this URL every 10 seconds until the video is ready, typically within 5 to 15 minutes. Use n8n's Loop or Wait node to handle this.

For Pika AI, extract 3 to 5 key medical concepts from your script. You can use Claude via API to pull these automatically.

Copy
POST https://api.pika.art/v1/videos/generation
Headers:
  Authorization: Bearer YOUR_PIKA_KEY
  Content-Type: application/json

Body:
{
  "prompt": "Medical educational animation: {{ extractedConcept }}. Professional style, clear visuals, suitable for patient education.",
  "aspect_ratio": "16:9",
  "duration": 10
}

Pika AI also returns a job_id for polling. Videos usually complete within 2 to 3 minutes.

Once you have the Hour One video and Pika AI visuals, you need to composite them. This is where your orchestrator becomes important. n8n doesn't natively handle video composition, so you have two options:

Option A: Use a Zapier-connected service like FFmpeg via an API wrapper (e.g., Mux, Cloudinary, or a custom Lambda function).

Option B: Use Claude Code within your n8n workflow to call FFmpeg locally if you're self-hosting.

For simplicity, consider uploading the Hour One video as your primary output and attaching Pika visuals as supplementary assets in your video description or as a YouTube playlist. Many clinics do this successfully without full frame-level composition.

Here's a practical approach: upload the Hour One video to YouTube using the YouTube Data API, then add timestamps and chapter links pointing to Pika AI visual sequences you've uploaded separately.

Copy
POST https://www.googleapis.com/youtube/v3/videos?part=snippet,status
Headers:
  Authorization: Bearer YOUR_YOUTUBE_TOKEN
  Content-Type: application/json

Body:
{
  "snippet": {
    "title": "{{ $json.topic }} - Patient Education",
    "description": "Learn about {{ $json.topic }}. [Visual aids: see description links below]",
    "tags": ["patient education", "healthcare", "{{ $json.topic }}"],
    "categoryId": "27"
  },
  "status": {
    "privacyStatus": "unlisted"
  }
}

Finally, update your Google Sheet with the completion status and video link.

Copy
{
  "resource": "spreadsheet",
  "operation": "updateRows",
  "spreadsheetId": "YOUR_SHEET_ID",
  "range": "Sheet1",
  "values": [
    {
      "row": "{{ $json.rowNumber }}",
      "videoUrl": "{{ $json.youtubeUrl }}",
      "status": "completed",
      "timestamp": "{{ now() }}"
    }
  ]
}

Setting Up in Zapier (Simpler)

Zapier's interface is visual, so you won't write code, but the logic is identical. Create a Zap with Google Sheets as the trigger.

Add a Zapier Code step to call ElevenLabs. Zapier allows Python or JavaScript.

Copyconst axios = require('axios');

const response = await axios.post(
  `https://api.elevenlabs.io/v1/text-to-speech/21m00Tcm4TlvDq8ikWAM`,
  {
    text: inputData.script,
    model_id: 'eleven_monolingual_v1',
    voice_settings: {
      stability: 0.5,
      similarity_boost: 0.75
    }
  },
  {
    headers: {
      'xi-api-key': process.env.ELEVENLABS_API_KEY,
      'Content-Type': 'application/json'
    }
  }
);

return { audioUrl: response.data.audio_url };

Add a Webhooks by Zapier action to call Hour One. Paste the same HTTP request structure shown above.

Use Zapier's Delay action to wait 60 seconds, then add a Webhooks by Zapier step to poll Hour One's video status endpoint until complete.

Connect to Pika AI similarly.

For video composition, Zapier integrates with Cloudinary. Use Cloudinary's API to overlay or composite your Pika visuals onto the Hour One video if frame-level precision matters.

Finally, update your Google Sheet with a built-in Zapier Google Sheets action.

Error Handling

Add error paths in both n8n and Zapier. If ElevenLabs fails, the workflow should log the error to a Slack channel and email the clinic staff. Use try-catch blocks in code steps.

Copytry {
  const response = await axios.post(apiUrl, payload);
  return { success: true, data: response.data };
} catch (error) {
  return {
    success: false,
    error: error.response?.data?.message || error.message,
    retryIn: 300 // 5 minutes
  };
}

Hour One and Pika AI both take time. If polling times out (e.g., video not ready after 30 minutes), set a maximum retry count and notify the clinic to check manually.

The Manual Alternative

If you prefer human control at certain steps, that's reasonable. You might automate voiceover and script generation but generate visuals manually. Or automate video creation but have a clinician review the script before it goes to ElevenLabs.

Insert a manual approval step in your orchestrator. In n8n, use the "Wait" node with a webhook; in Zapier, use the "Email" action to send a review link to a clinician. They approve or reject, which branches the workflow.

This adds friction but ensures medical accuracy. For high-stakes content like surgery explanations, this trade-off is worth it.

You can also generate multiple video versions (different narrators, avatars, lengths) and have staff select the best one before publishing. ElevenLabs and Hour One both support variants, making this practical.

Pro Tips

1. Manage API Rate Limits

ElevenLabs allows 10,000 characters monthly on the free tier, roughly 10 typical patient education scripts. Upgrade to a paid plan at $5 monthly for 100,000 characters if you plan more than a few videos per month.

Hour One's free tier limits video length to 2 minutes. Upgrade to $50 monthly for longer videos (up to 10 minutes).

Pika AI's free credits ($30 monthly value) generate roughly 2 to 3 full videos. Budget accordingly.

Use n8n's throttle node to space out API calls if you're processing many scripts at once, preventing rate limit hits.

Copy
{
  "waitTime": 2,
  "unit": "seconds"
}

2. Cache Voiceovers and Visuals

If your clinic reuses similar scripts (e.g., annual flu shot education with updated dates), cache the ElevenLabs audio. Store MP3 files in Google Drive or S3 and reference them instead of regenerating.

For Pika AI visuals, the same anatomical diagrams or process flows appear across multiple videos. Create a library and reference existing visuals; only generate new ones when the topic genuinely requires it.

3. Test with Short Scripts First

Run your first workflow with a 30-second script, not a 5-minute one. This confirms all API connections work and costs far less. Once you see a complete end-to-end video, scale up.

4. Monitor Voiceover Quality

ElevenLabs excels at natural speech, but medical terminology sometimes trips it up. Pre-test your script and add pronunciation hints.

Copy{
  "text": "The patient presents with myocardial infarction. [myˈkɑːrdiəl ɪnˌfɑːrkʃən]",
  "model_id": "eleven_multilingual_v2"
}

5. Version Control Your Scripts

Store all scripts in Google Sheets with timestamps. If a video needs revision, you can regenerate it from the original script. This matters when guidance changes (e.g., updated treatment protocols).

Add a column for script version and revision notes. Your workflow should reference the version, not just the latest row.

Cost Breakdown

Total estimated cost per month: £50–150 (approximately $60–180 USD) depending on video volume. At 10 videos monthly, that's £5–15 per video.

By automating patient education video creation, your clinic removes a major barrier to producing engaging health content. Staff enters a topic, the system handles the rest, and within an hour you have a polished, professional video ready for patients. No editing software. No video experience required. That's the power of chaining specialised AI tools together.