mirror of
https://github.com/khoaliber/n8nworkflows.xyz.git
synced 2026-04-28 00:29:22 +00:00
creation
This commit is contained in:
nusquama
+325
@@ -0,0 +1,325 @@
|
||||
Transcribe and Analyze Meetings or Podcasts with ElevenLabs and Google Gemini AI
|
||||
|
||||
https://n8nworkflows.xyz/workflows/transcribe-and-analyze-meetings-or-podcasts-with-elevenlabs-and-google-gemini-ai-8755
|
||||
|
||||
|
||||
# Transcribe and Analyze Meetings or Podcasts with ElevenLabs and Google Gemini AI
|
||||
|
||||
---
|
||||
|
||||
### 1. Workflow Overview
|
||||
|
||||
This workflow automates the transcription and analysis of audio or video content such as meetings, podcasts, or interviews using ElevenLabs’ Speech-to-Text API combined with Google Gemini AI for advanced transcript analysis. It is designed for scenarios requiring accurate speech transcription followed by insightful summarization and extraction of key information.
|
||||
|
||||
The workflow is composed of the following logical blocks:
|
||||
|
||||
- **1.1 Input Reception and Data Retrieval**: Fetches audio/video URLs from a Google Sheet to process.
|
||||
- **1.2 Speech-to-Text Processing**: Converts audio content to text transcripts via ElevenLabs API.
|
||||
- **1.3 Document Management**: Creates and updates Google Docs with transcripts.
|
||||
- **1.4 Transcript Analysis**: Uses Google Gemini AI to analyze and summarize the transcript.
|
||||
- **1.5 Output and Notification**: Converts analysis to HTML and sends an email summary.
|
||||
- **1.6 Batch Processing and Loop Control**: Manages batch processing of multiple audio URLs.
|
||||
- **1.7 Workflow Triggers**: Supports both manual and external workflow-triggered executions.
|
||||
|
||||
---
|
||||
|
||||
### 2. Block-by-Block Analysis
|
||||
|
||||
#### 2.1 Input Reception and Data Retrieval
|
||||
|
||||
**Overview:**
|
||||
This block initiates the workflow either manually or via another workflow and retrieves a list of audio/video URLs from a Google Sheet for processing.
|
||||
|
||||
**Nodes Involved:**
|
||||
- When clicking ‘Execute workflow’ (Manual Trigger)
|
||||
- When Executed by Another Workflow (Execute Workflow Trigger)
|
||||
- Get audio urls (Google Sheets)
|
||||
- Loop Over Items (Split In Batches)
|
||||
|
||||
**Node Details:**
|
||||
|
||||
- **When clicking ‘Execute workflow’**
|
||||
- *Type:* Manual Trigger
|
||||
- *Role:* Starts the workflow via user action in n8n UI.
|
||||
- *Connections:* Outputs to "Get audio urls".
|
||||
- *Failure modes:* None typical; manual start.
|
||||
|
||||
- **When Executed by Another Workflow**
|
||||
- *Type:* Execute Workflow Trigger
|
||||
- *Role:* Allows external workflows to trigger this workflow and pass a "text" parameter for direct analysis.
|
||||
- *Connections:* Outputs to "Transcript Analysis".
|
||||
- *Failure modes:* Missing or malformed input parameters.
|
||||
|
||||
- **Get audio urls**
|
||||
- *Type:* Google Sheets node
|
||||
- *Role:* Reads a Google Sheet to fetch rows containing audio URLs to transcribe.
|
||||
- *Configuration:*
|
||||
- Filters rows by the "DONE" column (likely to exclude already processed entries).
|
||||
- Reads from a specific Spreadsheet and sheet (gid=0).
|
||||
- *Credentials:* Google Sheets OAuth2.
|
||||
- *Outputs:* Passes data to "Loop Over Items".
|
||||
- *Failures:* Credential expiration, API limits, spreadsheet access issues.
|
||||
|
||||
- **Loop Over Items**
|
||||
- *Type:* Split In Batches
|
||||
- *Role:* Processes audio URLs one at a time to manage API rate limits and orderly execution.
|
||||
- *Connections:* Main output to "Speech-to-Text". Returns empty batch to "Speech-to-Text" for finalization.
|
||||
- *Failures:* Batch size misconfiguration or missing data.
|
||||
|
||||
---
|
||||
|
||||
#### 2.2 Speech-to-Text Processing
|
||||
|
||||
**Overview:**
|
||||
This block sends audio URLs to ElevenLabs API for transcription, handling conversion from audio/video to text.
|
||||
|
||||
**Nodes Involved:**
|
||||
- Speech-to-Text (HTTP Request)
|
||||
- Create Doc (Google Docs)
|
||||
- Add Transcript to Doc (Google Docs)
|
||||
- Update row (Google Sheets)
|
||||
- Call 'ElevenLabs Text-to-Speech' (Execute Workflow)
|
||||
|
||||
**Node Details:**
|
||||
|
||||
- **Speech-to-Text**
|
||||
- *Type:* HTTP Request
|
||||
- *Role:* Calls ElevenLabs Speech-to-Text API to transcribe audio from provided URLs.
|
||||
- *Configuration:*
|
||||
- POST request to `https://api.elevenlabs.io/v1/speech-to-text`.
|
||||
- Form-urlencoded body containing: model_id (`scribe_v1`), cloud_storage_url (dynamic from current item's AUDIO URL), tag_audio_events (`true`).
|
||||
- Authenticated by HTTP header with API key (`xi-api-key`).
|
||||
- *Inputs:* Audio URL from "Loop Over Items".
|
||||
- *Outputs:* JSON with transcription text.
|
||||
- *Failures:* Invalid API key, network timeout, unsupported audio format, API limits.
|
||||
|
||||
- **Create Doc**
|
||||
- *Type:* Google Docs
|
||||
- *Role:* Creates a new Google Doc to store the transcript.
|
||||
- *Configuration:*
|
||||
- Title set dynamically from transcription ID.
|
||||
- Specific Google Drive folder ID assigned.
|
||||
- *Credentials:* Google Docs OAuth2.
|
||||
- *Inputs:* Transcription metadata from "Speech-to-Text".
|
||||
- *Outputs:* Document ID used downstream.
|
||||
- *Failures:* Insufficient permissions, folder not found.
|
||||
|
||||
- **Add Transcript to Doc**
|
||||
- *Type:* Google Docs
|
||||
- *Role:* Inserts the transcript text into the newly created Google Doc.
|
||||
- *Configuration:* Inserts text from Speech-to-Text node into the document.
|
||||
- *Inputs:* Document ID from "Create Doc", transcript text.
|
||||
- *Failures:* Document access errors.
|
||||
|
||||
- **Update row**
|
||||
- *Type:* Google Sheets
|
||||
- *Role:* Marks the row as processed ("DONE" = "x") and stores the created Google Doc ID.
|
||||
- *Configuration:* Updates columns "DONE", "DOCS ID", using row_number for matching.
|
||||
- *Inputs:* Row number from batch, Doc ID from "Create Doc".
|
||||
- *Credentials:* Google Sheets OAuth2.
|
||||
- *Failures:* Sheet access errors or concurrent modification.
|
||||
|
||||
- **Call 'ElevenLabs Text-to-Speech'**
|
||||
- *Type:* Execute Workflow
|
||||
- *Role:* Invokes a sub-workflow (ID: 7igUJFLIXoP9aAKl) for text-to-speech conversion of the transcript text.
|
||||
- *Configuration:* Runs in "each" mode, waits for sub-workflow completion.
|
||||
- *Inputs:* Text from "Speech-to-Text".
|
||||
- *Failures:* Sub-workflow errors or misconfiguration.
|
||||
|
||||
---
|
||||
|
||||
#### 2.3 Transcript Analysis
|
||||
|
||||
**Overview:**
|
||||
This block processes the raw transcript text using Google Gemini AI (PaLM) to generate structured analysis and summaries.
|
||||
|
||||
**Nodes Involved:**
|
||||
- Transcript Analysis (Google Gemini AI)
|
||||
- Markdown (Markdown to HTML conversion)
|
||||
- Send email (Gmail)
|
||||
|
||||
**Node Details:**
|
||||
|
||||
- **Transcript Analysis**
|
||||
- *Type:* Google Gemini (PaLM) AI node
|
||||
- *Role:* Sends transcript text to Google Gemini model `models/gemini-2.5-pro` for in-depth analysis and summary.
|
||||
- *Configuration:*
|
||||
- System prompt defines detailed instructions for analyzing meetings, podcasts, interviews, extracting key points, decisions, action items, quotes, etc.
|
||||
- Input message includes full transcript text.
|
||||
- *Credentials:* Google Palm API (Gemini) OAuth2.
|
||||
- *Failures:* API quota limits, malformed input, response latency.
|
||||
|
||||
- **Markdown**
|
||||
- *Type:* Markdown converter
|
||||
- *Role:* Converts Gemini AI’s markdown-formatted summary into HTML for email readability.
|
||||
- *Inputs:* AI-generated content from "Transcript Analysis".
|
||||
- *Failures:* Formatting errors if AI output is malformed.
|
||||
|
||||
- **Send email**
|
||||
- *Type:* Gmail node
|
||||
- *Role:* Sends the final analyzed summary to a configured email address.
|
||||
- *Configuration:*
|
||||
- Recipient email set via parameter `YOUR_EMAIL_ADDRESS` (to be customized).
|
||||
- Email subject fixed as "=[N8N] Recap".
|
||||
- Sends plain text email with HTML content from "Markdown" node.
|
||||
- *Credentials:* Gmail OAuth2.
|
||||
- *Failures:* Authentication errors, quota limits, invalid recipient address.
|
||||
|
||||
---
|
||||
|
||||
#### 2.4 Batch Processing and Loop Control
|
||||
|
||||
**Overview:**
|
||||
Manages iterative processing of multiple audio URLs to ensure orderly execution and avoid API overload.
|
||||
|
||||
**Nodes Involved:**
|
||||
- Loop Over Items
|
||||
- Call 'ElevenLabs Text-to-Speech' (Execute Workflow)
|
||||
|
||||
**Node Details:**
|
||||
|
||||
- **Loop Over Items**
|
||||
- As previously described, splits input into manageable batches.
|
||||
|
||||
- **Call 'ElevenLabs Text-to-Speech'**
|
||||
- Invoked after updating the Google Sheet row to perform additional text-to-speech processing asynchronously for each transcript.
|
||||
|
||||
---
|
||||
|
||||
#### 2.5 Workflow Triggers
|
||||
|
||||
**Overview:**
|
||||
Allows two modes of triggering: manual execution and external workflow-based invocation.
|
||||
|
||||
**Nodes Involved:**
|
||||
- When clicking ‘Execute workflow’
|
||||
- When Executed by Another Workflow
|
||||
|
||||
**Node Details:**
|
||||
|
||||
- *Manual Trigger* is for testing or on-demand runs.
|
||||
- *Execute Workflow Trigger* allows integration with other automation pipelines, passing transcript texts directly for analysis.
|
||||
|
||||
---
|
||||
|
||||
### 3. Summary Table
|
||||
|
||||
| Node Name | Node Type | Functional Role | Input Node(s) | Output Node(s) | Sticky Note |
|
||||
|-------------------------------|-----------------------------|-------------------------------------------------|---------------------------------|-----------------------------------------|------------------------------------------------------------------------------------------------------------------|
|
||||
| When clicking ‘Execute workflow’ | Manual Trigger | Start workflow manually | — | Get audio urls | ## Transcribe and Analyze Meetings or Podcasts (audio/video) using ElevenLabs Speech-to-Text ... |
|
||||
| Get audio urls | Google Sheets | Retrieve audio URLs from Google Sheet | When clicking ‘Execute workflow’ | Loop Over Items | ## STEP 1 ... Register for FREE to [Elevenlabs](https://try.elevenlabs.io/ahkbf00hocnu) ... |
|
||||
| Loop Over Items | Split In Batches | Batch processing of audio URLs | Get audio urls | Speech-to-Text (main batch), Speech-to-Text (empty batch) | |
|
||||
| Speech-to-Text | HTTP Request | Transcribe audio via ElevenLabs API | Loop Over Items | Create Doc | |
|
||||
| Create Doc | Google Docs | Create Google Doc for transcript | Speech-to-Text | Add Transcript to Doc | |
|
||||
| Add Transcript to Doc | Google Docs | Insert transcript text into Google Doc | Create Doc | Update row | |
|
||||
| Update row | Google Sheets | Mark row as done and store Doc ID | Add Transcript to Doc | Call 'ElevenLabs Text-to-Speech' | |
|
||||
| Call 'ElevenLabs Text-to-Speech' | Execute Workflow | Sub-workflow call for text-to-speech conversion | Update row | Loop Over Items | |
|
||||
| When Executed by Another Workflow | Execute Workflow Trigger | Trigger workflow externally with transcript text | — | Transcript Analysis | ## STEP 2 ... Set YOUR_EMAIL_ADDRESS in "Send email" node |
|
||||
| Transcript Analysis | Google Gemini AI | Analyze and summarize transcript using AI | When Executed by Another Workflow | Markdown | |
|
||||
| Markdown | Markdown | Convert AI markdown summary to HTML | Transcript Analysis | Send email | |
|
||||
| Send email | Gmail | Send analyzed summary email | Markdown | — | |
|
||||
| Sticky Note | Sticky Note | Informational note | — | — | ## Transcribe and Analyze Meetings or Podcasts (audio/video) using ElevenLabs Speech-to-Text ... |
|
||||
| Sticky Note1 | Sticky Note | Instructions for setup and sheet cloning | — | — | ## STEP 1 ... Register for FREE to [Elevenlabs](https://try.elevenlabs.io/ahkbf00hocnu) ... |
|
||||
| Sticky Note2 | Sticky Note | Label for transcript section | — | — | ## Transcript |
|
||||
| Sticky Note3 | Sticky Note | Label for analysis section | — | — | ## Analysis |
|
||||
| Sticky Note4 | Sticky Note | Reminder to set email address | — | — | ## STEP 2 ... Set YOUR_EMAIL_ADDRESS in "Send email" node |
|
||||
|
||||
---
|
||||
|
||||
### 4. Reproducing the Workflow from Scratch
|
||||
|
||||
1. **Create a Manual Trigger Node**
|
||||
- Name: `When clicking ‘Execute workflow’`
|
||||
- Purpose: Allows manual start.
|
||||
|
||||
2. **Add a Google Sheets Node to Read Audio URLs**
|
||||
- Name: `Get audio urls`
|
||||
- Operation: Read rows from your specific Google Sheet (use your spreadsheet ID and sheet gid).
|
||||
- Filter: Exclude rows where "DONE" is marked.
|
||||
- Credentials: Configure Google Sheets OAuth2.
|
||||
|
||||
3. **Add a Split In Batches Node**
|
||||
- Name: `Loop Over Items`
|
||||
- Purpose: Process audio URLs sequentially or in manageable batches.
|
||||
|
||||
4. **Add HTTP Request Node for ElevenLabs Speech-to-Text**
|
||||
- Name: `Speech-to-Text`
|
||||
- Method: POST to `https://api.elevenlabs.io/v1/speech-to-text`
|
||||
- Content-Type: `application/x-www-form-urlencoded`
|
||||
- Body Parameters:
|
||||
- `model_id`: `scribe_v1`
|
||||
- `cloud_storage_url`: Set dynamically to current item’s AUDIO URL
|
||||
- `tag_audio_events`: `true`
|
||||
- Authentication: HTTP Header with key name `xi-api-key` and your ElevenLabs API key.
|
||||
|
||||
5. **Add Google Docs Node to Create Document**
|
||||
- Name: `Create Doc`
|
||||
- Title: Use a dynamic field (e.g., transcription ID or timestamp)
|
||||
- Folder: Specify your Google Drive folder ID
|
||||
- Credentials: Google Docs OAuth2.
|
||||
|
||||
6. **Add Google Docs Node to Insert Transcript Text**
|
||||
- Name: `Add Transcript to Doc`
|
||||
- Operation: Update document content by inserting the transcription text from the previous node.
|
||||
|
||||
7. **Add Google Sheets Node to Update the Processed Row**
|
||||
- Name: `Update row`
|
||||
- Operation: Mark "DONE" with "x", store the created Google Doc ID, match by row_number.
|
||||
- Credentials: Google Sheets OAuth2.
|
||||
|
||||
8. **Add Execute Workflow Node to Call Sub-Workflow for Text-to-Speech**
|
||||
- Name: `Call 'ElevenLabs Text-to-Speech'`
|
||||
- Configuration: Pass the transcript text as input parameter `text`.
|
||||
- Mode: Each (process items individually), wait for sub-workflow completion.
|
||||
- Workflow ID: Use your sub-workflow's ID for text-to-speech.
|
||||
|
||||
9. **Establish Connections:**
|
||||
- Connect manual trigger → Google Sheets (Get audio urls) → Split In Batches (Loop Over Items) → Speech-to-Text → Create Doc → Add Transcript to Doc → Update row → Call 'ElevenLabs Text-to-Speech' → Loop Over Items (for next batch).
|
||||
|
||||
10. **Add Execute Workflow Trigger Node**
|
||||
- Name: `When Executed by Another Workflow`
|
||||
- Accept input parameter named `text`.
|
||||
- Connect output to `Transcript Analysis`.
|
||||
|
||||
11. **Add Google Gemini AI Node**
|
||||
- Name: `Transcript Analysis`
|
||||
- Model: `models/gemini-2.5-pro`
|
||||
- System Message: Use the provided detailed prompt for meeting/podcast transcript analysis.
|
||||
- Input Message: Full transcript text from input.
|
||||
- Credentials: Google Palm API OAuth2.
|
||||
|
||||
12. **Add Markdown Node**
|
||||
- Name: `Markdown`
|
||||
- Mode: Markdown to HTML conversion.
|
||||
- Input: Output from `Transcript Analysis`.
|
||||
|
||||
13. **Add Gmail Node**
|
||||
- Name: `Send email`
|
||||
- Recipient: Set your email address in the sendTo parameter.
|
||||
- Subject: "=[N8N] Recap"
|
||||
- Message: Use HTML content from Markdown node.
|
||||
- Credentials: Gmail OAuth2.
|
||||
|
||||
14. **Connect `When Executed by Another Workflow` → `Transcript Analysis` → `Markdown` → `Send email`.**
|
||||
|
||||
15. **Add Sticky Notes for Documentation** (Optional but recommended)
|
||||
- Add notes describing setup steps, usage instructions, and labeling sections.
|
||||
|
||||
---
|
||||
|
||||
### 5. General Notes & Resources
|
||||
|
||||
| Note Content | Context or Link |
|
||||
|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------|
|
||||
| This workflow supports transcription of audio/video files up to 3 GB and 10 hours duration using ElevenLabs API. It supports multilingual recognition across 99 languages. | ElevenLabs Speech-to-Text API official features. |
|
||||
| Register for a free ElevenLabs account at [ElevenLabs](https://try.elevenlabs.io/ahkbf00hocnu). Clone the sample Google Sheet for input data at [Google Sheet Clone](https://docs.google.com/spreadsheets/d/1046e1OLfwiu1vfvnX_i2LP3yXOGL_87_BxyiXgkVmuU/edit?usp=sharing). | Setup Instructions Sticky Note. |
|
||||
| Set the HTTP header authentication in the Speech-to-Text node with header name `xi-api-key` and your ElevenLabs API key. | Credential configuration details. |
|
||||
| Customize the email recipient address in the “Send email” node to receive the transcript analysis summaries. | Reminder Sticky Note. |
|
||||
| The Google Gemini AI system prompt is crafted to extract detailed meeting insights including decisions, action items, participants, and next steps. It uses structured markdown formatting for clarity and readability. | AI prompt customization for transcript analysis. |
|
||||
|
||||
---
|
||||
|
||||
**Disclaimer:**
|
||||
The text and data processed by this workflow are handled strictly according to current content policies and contain no illegal or offensive elements. All data is legal and publicly accessible.
|
||||
|
||||
---
|
||||
Reference in New Issue
Block a user