Transcribe monologue-style audio into structured CRM records using n8n, OpenAI (GPT-3.5/4o), and Airtable. This hybrid-code automation extracts caller intent, urgency, deadlines, and follow-up actions into clean, ready-to-use fields.
- 🎙️ Whisper-based audio transcription
- 🧠 GPT-powered LLM field extraction
- ⏳ Deadline & urgency detection
- 🧹 Fallback logic (for missing fields)
- 📤 Airtable CRM integration
- 📝 Summary generation for transcripts
Version | Description |
---|---|
v0.1 | Regex-only transcription → Google Sheets |
v0.2 | Hugging Face model added |
v0.3 | Switched to OpenAI API (GPT-3.5) |
v1.0 | Airtable CRM version with fallback logic, deadline/date extraction, and GPT-4o compatibility |
See /sample-output
for actual JSONs from Camille, Marcus, and Samantha test cases.
👉 View Airtable CRM Demo Sheet
<iframe class="airtable-embed" src="https://airtable.com/embed/appXpNiN4xZXPi7kr/shr4M3RUqg7cJ4Asc?viewControls=on" frameborder="0" onmousewheel="" width="100%" height="533" style="background: transparent; border: 1px solid #ccc;"></iframe>
- Import the
transcription_workflow_airtable_crm_v1.json
into your n8n instance - Link your OpenAI, Airtable, and Telegram (optional) credentials
- Feed audio via URL or manual trigger
- Review outputs directly in Airtable
MIT License — see LICENSE
for full terms.