WhatsApp AI Command Center

Who is this for?

You lead support triaging 40 WhatsApp messages with voice, photos and PDFs a day, and this WhatsApp MongoDB workflow transcribes and drafts from docs. Each message needs audio played, image viewed, and file opened with a reply crafted.

What problem does this solve?

You play voice notes, view photos, then open PDFs and docs to craft replies. When a new WhatsApp message arrives, the flow transcribes audio, describes images, extracts files, and drafts from the MongoDB base. You just approve edge cases instead of typing every reply.

What this workflow does

1. Download Image (HTTP API)
2. Analyze Image (OpenAI)
3. Download Audio (HTTP API)
4. Transcribe Audio (OpenAI)

How it works

A new WhatsApp message kicks it off → Input type, Text, AI Agent1, ending in Fix mimeType for Audio.

Setup

1. Open danielgrey.shop/app/templates/dg-0010 → Copy JSON → n8n → Import from Clipboard (or Push to n8n)
2. Connect credentials: Google Docs, HTTP Header Auth, MongoDB, OpenAI, WhatsApp
3. Run once manually to verify each branch — it imports clean and goes green first try — then a 10-minute review and activate.

Requirements

- n8n 1.50.0 or later
- Credentials for Google Docs, HTTP Header Auth, MongoDB, OpenAI, WhatsApp
- One-time $199 — yours forever, Pro unlocks it too.

Customising

- Tune the logic inside Fix mimeType for Audio — thresholds and weights are plain code with comments.
- Adjust the branch condition in Only PDF File to match your own cutoff.