AI Voice Note Processing and Archiving System - n8n Workflow

Knowledge Managers & Researchers: Individuals who frequently record thoughts or meetings and need them summarized, structured, and archived immediately.
Automation Engineers: Users looking for advanced examples of using the n8n AI Agent node with external LLM providers like OpenRouter.
Team Leaders: Those needing a robust n8n workflow solution to turn raw audio inputs into structured, accessible documentation.
Tech Teams utilizing NocoDB and Nuclino: Users integrating various tools into a unified knowledge capture system using n8n.

Manually transcribing and summarizing voice notes, followed by documenting them across multiple platforms, is a time-consuming administrative task. This specific n8n workflow solves this by creating a fully automated pipeline. The solution starts with an external voice note transcription service triggering an initial n8n trigger. The transcribed prompt is then fed into an AI Agent powered by Claude Sonnet via OpenRouter, which generates a highly structured Markdown response. The resulting prompt and response data are simultaneously archived in NocoDB for historical record-keeping and published to Nuclino as a permanent knowledge base entry. Finally, a notification is dispatched via Slack, ensuring the team is immediately aware of the new, summarized insight. This use of an n8n node combination streamlines operational efficiency and ensures knowledge capture is instantaneous and complete.

This automation begins with a Webhook n8n trigger, which listens for a POST request containing the voice note transcription (transcript) and title.

1. Ingestion and Preparation: The Webhook passes the payload to the Edit Fields n8n node, which renames the transcript to prompt and extracts the note_title.


2. AI Analysis: The AI Agent n8n node receives the user's prompt. It is configured with a complex system message forcing the response into a structured Markdown format (including SUMMARY, PROMPT, and RESPONSE sections). This agent leverages the OpenRouter Chat Model node, which uses the Anthropic Claude Sonnet 4 LLM.


3. Data Consolidation: The Merge n8n node combines the original prompt data (from Edit Fields) with the detailed AI output (from AI Agent).


4. Data Restructuring: The Edit Fields1 n8n node prepares the consolidated data for archival and publishing, ensuring all key fields (prompt, output, title, creation timestamp) are correctly mapped.


5. Archival (NocoDB): Two sequential NocoDb n8n node instances (Save Prompt and Save Output) archive the original input and the final AI output into separate tables for structured record-keeping.


6. Documentation (Nuclino): The HTTP Request n8n node creates a new item in Nuclino. The content is meticulously formatted using Markdown, embedding the note title, creation date, original prompt, and the detailed AI response.


7. Notification (Slack): The final Send a message Slack n8n node broadcasts a rich summary message, including the prompt, the AI response, and the direct link to the newly created Nuclino document, completing this comprehensive n8n workflow.

To deploy this powerful n8n workflow template, follow these steps:

1. Import: Copy the provided JSON code and import it into your n8n instance via the 'Workflows' > 'New' > 'Import from JSON' option.


2. Webhook Setup: Activate the Webhook n8n trigger node. Copy the production URL and configure your voice note transcription service (e.g., Voicenotes.com, if applicable) to send POST requests containing the title and transcript to this endpoint.


3. Credential Setup:

1. Activation: Once all credentials and IDs are configured, set the n8n workflow to 'Active'.

This n8n workflow utilizes several specialized nodes to achieve its goals:

Webhook (n8n trigger): The primary entry point, configured to listen for POST requests at a specific path, initiating the entire n8n workflow upon receiving a new voice note transcription.
Edit Fields (Set n8n node): Used twice (Edit Fields and Edit Fields1) to map and rename incoming JSON fields. Key Configuration: Maps body.data.transcript to prompt and body.data.title to note_title.
AI Agent (n8n node): Executes the core analytical logic. Key Configuration: Uses a highly prescriptive system message to ensure the output is delivered in strict Markdown format, including specific sections (SUMMARY, PROMPT, RESPONSE).
OpenRouter Chat Model (n8n node): Provides the large language model resource for the AI Agent, configured specifically to use the anthropic/claude-sonnet-4 model.
Merge (n8n node): Combines data items from two separate branches (input context and AI output) to ensure all required fields are available downstream.
NocoDb (n8n node): Used for persistent data storage and archiving. Key Configuration: Configured in 'create' operation mode, saving data to specific tables based on field name/value mappings.
HTTP Request (n8n node): Interacts with the Nuclino API to create documentation. Key Configuration: Uses a POST method, sending a custom Markdown body structure containing the prompt and AI response to the /v0/items endpoint.
Send a message (Slack n8n node): Distributes the final result to a designated Slack channel, formatted with markdown (mrkdwn: true) for readability.