AI · Full-Stack · Case study
LeadCall AI
Multi-tenant outbound-calling SaaS built on the voice-agent platform. Upload a lead list and a pitch document, and Celery workers dial through the platform API, sync transcripts back and keep each tenant’s leads and credentials apart.
- Stack
- Python · FastAPI · Celery · Alembic · PostgreSQL · React · RAG · Docker Compose
Situation
Small teams sit on lead lists they never call. The blocker is not the conversation. It is the operations around it: dialing at the right pace, knowing which calls actually finished, and keeping one client’s data away from another’s.
Task
Build a service where uploading a lead list and a pitch document is enough to run a calling campaign, with each tenant’s leads, credentials and call history kept separate.
Action
LeadCall AI is FastAPI + PostgreSQL with Alembic migrations and a React dashboard. It does not run its own telephony: it dials through the CortexCraft voice-agent platform’s external API, and each tenant stores its own platform URL and an encrypted API key. Celery workers handle dialing and webhook callbacks, leads arrive by Excel import, and pitch documents are cleared and re-indexed for RAG so the agent’s script matches what the client sells. Transcripts sync back per call, and the call log exports for the client.
Result
A lead list and a pitch document become a running campaign with per-call transcripts and an exportable call log, without the client operating any telephony of their own.
Multi-tenant
encrypted per-tenant keys
Celery
dial + callback workers
Excel
lead import + log export