Automated Sales Call Transcriptions & Intelligent CRM Analysis powered by Gemini & Google Cloud
of a sales rep's valuable weekly working hours.
Upload any audio recording or text document to immediately process, extract, and populate all vital metrics.
Convert audio to text & map speakers.
Gemini extracts key insights & scores.
Editable CRM card ready for review.
Clean HTML5, CSS3, & Vanilla JS. Fast, lightweight, reactive Firestore listeners for real-time status updates.
Firebase Cloud Functions running Node.js 20. Configured with 2GB RAM to transcode audio files using FFMPEG instantly.
Google Speech-to-Text v2 for diarization. Gemini 2.5 Pro via Vertex AI ensuring 100% structured JSON outputs.
Direct HTTP functions fail on large media uploads.
✓ Solution: Client-side Firebase Storage SDK uploads, then backend triggers local fast download.
Google STT returns Speaker 1 & 2 without identifying roles.
✓ Solution: Gemini analyzes context to match Speaker IDs to "Manager" vs "Client".
NPM packages (like pdf-parse) changed their API causing 500 errors.
✓ Solution: Locked functions down to stable, verified package versions (v1.1.1).
REST API · Webhooks · MCP Protocol · Zapier · Make