[Example] 460 — Webex Recording Transcription (Node.js) (#177)

## New example: Webex Recording Transcription

<!-- metadata
type: example
number: 460
slug: webex-recording-transcription-node
language: Node.js
products: stt
integrations: webex
-->

**Integration:** Cisco Webex | **Language:** Node.js | **Products:** STT

### What this shows
A Node.js Express server that listens for Webex `meetingRecording.ready`
webhooks, downloads the meeting recording audio via the Webex REST API,
and transcribes it using Deepgram nova-3 with speaker diarization and
smart formatting. Optionally posts the transcript back to a Webex space.

### Research note
The original issue (#137) requested a real-time audio streaming bot
using webex-node-bot-framework. Research found that:
- webex-node-bot-framework is a **messaging bot** framework — it cannot
join meetings or access audio streams
- Webex does **not** expose real-time meeting audio via public APIs
- The feasible approach is **recording-based transcription** using the
Recordings API + webhooks

This example implements the recording-based approach, which is the
practical way to transcribe Webex meetings.

### Required secrets
- `DEEPGRAM_API_KEY`
- `WEBEX_BOT_TOKEN` — from Webex Developer Portal
- `WEBEX_WEBHOOK_SECRET` — user-defined shared secret for webhook
verification

### Tests
✅ Tests passed (with test credentials for webhook verification)
⚠️ Full end-to-end test requires real Webex credentials

```
Test 1: createApp() returns a configured Express app...
✓ createApp() returned an Express app

Test 2: GET /health returns { status: "ok" }...
✓ GET /health returned { status: "ok" }

Test 3: POST /webhook — invalid signature returns 401...
✓ Invalid signature correctly rejected with 401

Test 4: POST /webhook — non-recording event returns ignored...
✓ Non-recording event correctly returned { status: "ignored" }

Test 5: POST /webhook — valid recording.ready returns processing...
✓ Valid recording.ready event accepted with { status: "processing" }

✓ All tests passed
```

Closes #137

---
*Built by Engineer on 2026-04-04*

Co-authored-by: examples-bot <noreply@deepgram.com>
Co-authored-by: Claude <noreply@anthropic.com>

Author: 3 people

examples/460-webex-recording-transcription-node/.env.example

@@ -0,0 +1,6 @@
+# Deepgram — https://console.deepgram.com/
+DEEPGRAM_API_KEY=
+
+# Webex — https://developer.webex.com/my-apps
+WEBEX_BOT_TOKEN=
+WEBEX_WEBHOOK_SECRET=

examples/460-webex-recording-transcription-node/.npmrc

@@ -0,0 +1 @@
+save-exact=true

examples/460-webex-recording-transcription-node/README.md

@@ -0,0 +1,71 @@
+# Webex Recording Transcription with Deepgram
+
+Automatically transcribe Cisco Webex meeting recordings using Deepgram's nova-3 speech-to-text model. When a Webex meeting recording becomes available, this server receives a webhook, downloads the audio, transcribes it with Deepgram, and optionally posts the transcript back to a Webex space.
+
+## What you'll build
+
+A Node.js Express server that listens for Webex `meetingRecording.ready` webhooks, downloads the recording audio via the Webex REST API, sends it to Deepgram for transcription with speaker diarization, and logs the formatted transcript.
+
+## Prerequisites
+
+- Node.js 18+
+- pnpm
+- Deepgram account — [get a free API key](https://console.deepgram.com/)
+- Webex account — [create a bot](https://developer.webex.com/my-apps/new/bot)
+
+## Environment variables
+
+| Variable | Where to find it |
+|----------|-----------------|
+| `DEEPGRAM_API_KEY` | [Deepgram console](https://console.deepgram.com/) |
+| `WEBEX_BOT_TOKEN` | [Webex Developer Portal → My Apps](https://developer.webex.com/my-apps) — copy the Bot Access Token |
+| `WEBEX_WEBHOOK_SECRET` | You choose this value when creating the webhook via the Webex API |
+
+## Install and run
+
+```bash
+cp .env.example .env
+# Fill in your credentials in .env
+
+pnpm install
+pnpm start
+```
+
+Then register a Webex webhook pointing to your server:
+
+```bash
+curl -X POST https://webexapis.com/v1/webhooks \
+  -H "Authorization: Bearer $WEBEX_BOT_TOKEN" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "name": "Recording Transcription",
+    "targetUrl": "https://your-server.example.com/webhook",
+    "resource": "meetingRecordings",
+    "event": "ready",
+    "secret": "your-webhook-secret"
+  }'
+```
+
+## Key parameters
+
+| Parameter | Value | Description |
+|-----------|-------|-------------|
+| `model` | `nova-3` | Deepgram's most accurate general-purpose model |
+| `diarize` | `true` | Enables speaker labels for multi-speaker meetings |
+| `smart_format` | `true` | Adds punctuation, capitalization, and number formatting |
+| `paragraphs` | `true` | Groups transcript into readable paragraphs |
+
+## How it works
+
+1. A Webex meeting ends and the recording is processed by Webex
+2. Webex sends a `meetingRecording.ready` webhook to this server
+3. The server verifies the webhook signature (HMAC-SHA1)
+4. It fetches the recording metadata from the Webex Recordings API
+5. It downloads the audio file using the temporary direct download link
+6. The audio buffer is sent to Deepgram's pre-recorded transcription API
+7. The transcript is logged, with speaker labels and paragraph formatting
+8. Optionally, the transcript is posted back to a Webex space
+
+## Starter templates
+
+[deepgram-starters](https://github.com/orgs/deepgram-starters/repositories)

examples/460-webex-recording-transcription-node/package.json

@@ -0,0 +1,24 @@
+{
+  "name": "webex-recording-transcription-node",
+  "version": "1.0.0",
+  "description": "Transcribe Webex meeting recordings using Deepgram nova-3",
+  "main": "src/server.js",
+  "packageManager": "pnpm@9.6.0",
+  "scripts": {
+    "start": "node src/server.js",
+    "test": "node tests/test.js"
+  },
+  "dependencies": {
+    "@deepgram/sdk": "5.0.0",
+    "dotenv": "16.4.7",
+    "express": "4.22.1"
+  },
+  "pnpm": {
+    "overrides": {
+      "path-to-regexp": "0.1.13"
+    }
+  },
+  "engines": {
+    "node": ">=18"
+  }
+}