SheetPulse MMS AI is a production-shaped MMS automation control plane for teams that want:
- Google Sheets as the operator workspace
- Twilio as the MMS delivery rail
- AI-assisted personalization and launch previews
- consent-aware, auditable dispatch rather than blind blasting
It is designed for high-throughput opted-in messaging, not for bypassing Twilio, carrier, or compliance controls. It does not promise unlimited or policy-evasive sending.
- FastAPI API for campaign preview, launch simulation, Sheets contact viewing, and Twilio callbacks
- Google Sheets ingestion layer with service-account support and demo fallback
- Twilio MMS orchestration layer with simulated and live modes
- AI personalization layer using the OpenAI Responses API when credentials are present
- React operator dashboard for compliance, lane status, preview copy, and decisions
- tests, browser smoke, GitHub Actions CI, screenshots, and detailed docs
sheetpulse-mms-ai/
|-- apps/
| |-- api/
| `-- web/
|-- docs/
|-- fixtures/
|-- src/sheetpulse_mms/
`-- tests/
python -m pip install -r requirements.txtnpm run install:webpython -m uvicorn apps.api.app.main:app --host 127.0.0.1 --port 8020npm run dev --prefix apps/webSHEETPULSE_DEMO_MODE=1GOOGLE_SERVICE_ACCOUNT_JSON=...GOOGLE_SHEETS_SPREADSHEET_ID=...GOOGLE_SHEETS_RANGE=Contacts!A1:Z5000TWILIO_ACCOUNT_SID=...TWILIO_AUTH_TOKEN=...TWILIO_MESSAGING_SERVICE_SID=...TWILIO_FROM_NUMBER=...OPENAI_API_KEY=...OPENAI_MODEL=gpt-4.1-mini
The original request asked for “without any limits” and “unlimited MMS,” but this repo intentionally does not implement limit evasion. Twilio, carriers, and opt-in law all impose real constraints. So the platform is built around:
- opt-in enforcement
- quiet-hour suppression
- segment targeting
- preview and simulation before launch
- clear callback and audit surfaces
That keeps it useful for legitimate business operations while staying production-realistic.
- Twilio Messages resource
- Twilio Message Scheduling
- Twilio Messaging Services
- Google Sheets API quickstart
- OpenAI Responses API
The repo was verified with:
python -m pytest tests -qnpm run typechecknpm run build:web- live API smoke
- live browser smoke and screenshot capture
More detail is in docs/final-audit.md.


