SingleTable | Universal Restaurant Booking Sync - Sell Project

​ SingleTable: AI Booking Automation for TheFork, Gmail and WhatsApp - trained on real data. SingleTable is a full-stack restaurant booking automation codebase combining AI email interpretation, structured booking records, persistent task management, Playwright browser automation and WhatsApp notifications. It is designed to turn incoming reservation emails into actions in TheFork Manager, reducing the repeated work of reading emails, identifying changes and manually updating reservations. Technical architecture • AI-powered email extraction Google Gemini interprets booking emails and extracts structured JSON containing the booking source, reservation reference, guest details, party size, date, time and notes. The extraction logic distinguishes new bookings, modifications, cancellations and enquiries, while filtering unrelated messages. • Incremental Gmail ingestion The Gmail integration uses OAuth 2.0 with read-only access. Saved processing timestamps support incremental retrieval across scheduled runs, avoiding the need to scan the entire mailbox each time. • Structured validation and booking states Python and Pydantic models organise extracted information into a consistent booking schema. Processing logic checks booking dates and event types before creating records and execution tasks. • Persistent workflow management Google Cloud Firestore stores booking records, pending actions, execution results, timestamps and failure details. Email processing and browser execution are separate stages, allowing an action to be tracked from ingestion through execution. • Playwright browser automation A Node.js automation layer connects to Chrome through the Chrome DevTools Protocol and operates within an authenticated TheFork Manager session. The implementation combines browser interaction with browser-context requests for reservation operations. • Duplicate-booking checks Before creating a reservation, the browser worker checks existing bookings using guest-name, date and time information. Detected matches are recorded as skipped duplicates. • Restaurant-specific configuration The codebase includes restaurant-scoped Firestore paths, configuration handling and dedicated Chrome profile paths. An orchestration layer coordinates processing for configured restaurants. • Notifications and operational visibility Twilio integration provides WhatsApp booking notifications, processing summaries and failure alerts. Task records retain execution outcomes and error information for troubleshooting. Technology stack Backend: Python, FastAPI and Pydantic Automation: Node.js, Playwright and Chrome/Chromium Email: Gmail API and Google OAuth 2.0 AI extraction: Google Gemini Persistence: Google Cloud Firestore Messaging: Twilio WhatsApp Analytics: BigQuery synchronisation scripts Codebase components Python email-ingestion and booking-processing modules, FastAPI backend, booking models, Firestore helpers, TheFork automation adapter, notification integrations, orchestration scripts and deployment documentation. Intended buyers Developers, automation agencies and hospitality technology teams implementing restaurant booking workflows or adapting the components for related operational processes. Setup requirements A compatible hosting environment, Google Cloud configuration, Gmail OAuth credentials, Gemini API access, Twilio WhatsApp configuration and an authorised TheFork Manager account are required. Restaurant-specific settings and initial browser-session setup must be configured before operation. Third-party accounts and usage charges are separate.