How to automate meeting notes with AI - from transcript to assigned tasks

You can automate meeting notes with AI by feeding an audio recording into a transcription service, passing the raw text to a Large Language Model (LLM) that extracts a concise summary and a list of action items, then routing those items into Notion for reference and Asana for execution - all orchestrated in n8n. The result is a hands-free workflow that turns every meeting into a searchable knowledge base and a ready-to-act task list without manual copy-pasting.

What you need

Estimated build time: 2-3 hours for a first-pass workflow, plus 30 minutes for testing and tweaking.

Step-by-step build

1. Capture audio and get a transcript 1. Sign up for Fireflies.ai and install the Chrome/Zoom integration. 2. In your meeting, let Fireflies join as a participant (ff.ai). After the call ends, Fireflies will email you a link to the raw transcript (plain-text). 3. Enable the Webhooks option on the transcript page (Settings → Integrations → Webhook). Set the target URL to the n8n webhook you'll create next.

2. Create an n8n webhook trigger 1. Deploy n8n locally with Docker:

 docker run -d --name n8n \
 -p 5678:5678 \
 -e N8N_BASIC_AUTH_ACTIVE=true \
 -e N8N_BASIC_AUTH_USER=admin \
 -e N8N_BASIC_AUTH_PASSWORD=YOUR_PASSWORD \
 n8nio/n8n
 

1. In the n8n UI, click + New Workflow, add a Webhook node, set HTTP Method to POST, and copy the generated URL (e.g., https://your-host.com/webhook/meeting-notes).
2. Save the workflow - this URL is what you paste into Fireflies' webhook field.

3. Send transcript to OpenAI for summary & action extraction 1. Add an HTTP Request node after the webhook. 2. Set Method to POST, URL to https://api.openai.com/v1/chat/completions. 3. In Authentication, choose Header Auth and add Authorization: Bearer {{ $env.OPENAI_API_KEY }} (store the key in n8n's Credentials → API Key).

1. Use the following JSON body (this is the prompt that extracts both a summary and bullet-point actions):

 {
 "model": "gpt-4o-mini",
 "temperature": 0,
 "messages": [
 {
 "role": "system",
 "content": "You are an assistant that turns meeting transcripts into a short summary and a list of actionable items. Return JSON with two keys: summary (max 3 sentences) and actions (array of objects with title and assignee if mentioned)."
 },
 {
 "role": "user",
 "content": "{{ $json.body.transcript }}"
 }
 ]
 }
 

What this does: Sends the raw transcript to OpenAI, asking the model to output a deterministic JSON object containing a concise meeting summary and any identified action items.

1. Add a Set node to parse {{ $json.choices[0].message.content }} into two fields: summary and actions. Use an expression like {{$json["summary"]}} after you JSON.parse the string.

4. Push the summary to Notion 1. Drag a Notion node, authenticate via OAuth (n8n has a built-in Notion credential). 2. Choose Create Page in the database you prepared for meeting notes. 3. Map fields: Title → `Meeting - {{ $json.meetingDate }}` (extract date from webhook payload) Properties → Summary → {{$json.summary}} Content* → {{ $json.actions | json }} (store actions as raw JSON for later reference)

5. Create tasks in Asana 1. Add an Asana node, connect with a Personal Access Token (PAT). 2. Set Operation to Create Task. 3. Loop over the actions array using an IF node + SplitInBatches (batch size = 1). For each action: Name → `{{ $json.title }}` Assignee → {{ $json.assignee || "" }} (if the model identified a name, you may need a lookup table to map to Asana user IDs) Notes → `{{ $json.summary }}` (provides context) Project → Your "Meeting Action Items" project ID

6. Optional: Notify the team on Slack 1. Add a Slack node (OAuth token). 2. Send a message to a channel like #meeting-recap with a formatted block:

 *Meeting recap:* {{ $json.summary }}
 *Action items:* {{ $json.actions | json }}
 

7. Test end-to-end 1. Run a short Zoom call, let Fireflies record, then check the n8n execution log. 2. Verify that Notion contains a new page and Asana shows tasks with correct assignees. 3. Tweak the system prompt (step 3) if the model misses nuances (e.g., "When no assignee is named, assign to the meeting host").

Using GPT-4's 8,192-token context window, a 30-minute transcript (~9 k words ≈ 13 k tokens) fits comfortably, guaranteeing full-text analysis without truncation.

Where this breaks

For a deeper technical reference, see n8n's documentation.