PayPulse — Agentic Treasury & Autonomous Procurement OS

Built for the PayPal AI Hackathon ($67,500 Prize Pool across 23 Winners)
Live Interactive Command Center: https://jeffrey-graham.github.io/paypulse-paypal-ai-agent/

PayPulse Architecture

🎯 Target Tracks & Multi-Sponsor Alignment

PayPulse is engineered from the ground up as a unified autonomous B2B procurement and treasury operating system that qualifies for the Main Track (Best PayPal + AI & Best Agentic Commerce) and all 4 Multi-Winner Sponsor Cash Tracks:

  1. Main Track — Best PayPal + AI & Best Agentic Commerce ($12k / $8k / $5k + $5k + $5k):
    • Autonomous AI Procurement Agent exposing MCP/OpenAI-compatible tool definitions (channel3_search_and_rank_quotes, paypal_autonomous_procure_and_schedule, paypal_execute_vendor_payout_batch) backed by PayPal Orders v2, PayPal Invoicing v2, PayPal Payouts v1, and cryptographic Webhook Signature Verification (CRC32 + HMAC-SHA256).
  2. Best Use of AG Grid ($5,000 1st / $2,000 2nd / $1,000 x 3 3rd):
    • Live AG Grid Enterprise Financial Ledger & 3-Way Match Dispute Grid (AGGridLedgerReconciler + live AG Grid CDN in index.html) reconciling Channel3 Purchase Orders against Supplier Invoices and PayPal Captures with instant sorting, quick-filtering, custom variance badges, CSV export, and 1-click dispute resolution.
  3. Best Use of APIMatic ($1,000 x 3 Winners):
    • Strongly-typed APIMatic PayPal SDK Client Layer (src/apimatic_paypal_sdk.py) enforcing OpenAPI schema validation (Money, OrderItem, PurchaseUnitRequest, OrderRequest, InvoiceDraftRequest, PayoutBatchRequest), proactive OAuth 2.0 /v1/oauth2/token lifecycle caching, and deterministic PayPal-Request-Id Idempotency Middleware that blocks duplicate captures and detects payload conflicts.
  4. Best Use of Bryntum ($1,000 x 3 Winners):
    • Bryntum Settlement Milestone Gantt & Multi-Vendor Escrow Scheduler (BryntumSettlementScheduler + interactive Gantt timeline in index.html) modeling Finish-to-Start (FS) dependencies (M1: 30% Order Auth Deposit ➔ M2: Carrier Transit & Telemetry QA ➔ M3: 3-Way Match & 70% PayPal Capture/Payout) and 30-day Treasury Liquidity forecasting.
  5. Best Use of Channel3 ($1,500):
    • Channel3 AI Product Discovery & Price Intelligence Client (Channel3CatalogClient) querying multi-merchant AI catalogs, applying volume tier discounts, calculating landed unit economics, and ranking vendor quotes by a multi-factor procurement utility score (50% Price Competitiveness + 30% Merchant Reliability + 20% Lead-Time SLA).

🏗️ Repository Structure

paypal-ai-hackathon/
├── index.html                      # Interactive PayPulse Command Center (AG Grid + Bryntum + Channel3 + APIMatic)
├── src/
│   ├── __init__.py
│   ├── apimatic_paypal_sdk.py      # APIMatic Strongly-Typed PayPal SDK, OAuth2, Idempotency & Webhook Verifier
│   └── paypulse_engine.py          # Channel3 Catalog Client, AG Grid 3-Way Reconciler, Bryntum Scheduler & Agentic Tools
├── tests/
│   └── test_paypulse.py            # 8 End-to-End Unit Tests (100% Pass)
├── scripts/
│   └── generate_demo_video.py      # Generates 1080p assets/architecture.png & assets/paypal_ai_demo.mp4
├── assets/
│   ├── architecture.png            # 1080p System Architecture Diagram
│   └── paypal_ai_demo.mp4          # 1080p H.264 Walkthrough Video
├── DEVPOST_SUBMISSION.md           # Complete Devpost Submission Copy
└── README.md

🚀 Quickstart & Verification

1. Run the Unit Test Suite

PYTHONPATH=. python3 -m unittest discover -s tests -v

All 8 tests verify APIMatic schema validation, OAuth2 token lifecycle, PayPal-Request-Id idempotency replay & conflict detection, webhook CRC32 + HMAC-SHA256 verification, Channel3 quote ranking, 3-Way Match auto-capture vs. dispute freeze (>1.5% variance), and Bryntum Finish-to-Start milestone scheduling.

2. Launch the Interactive Command Center

Open index.html in any browser or visit the live deployment:
👉 https://jeffrey-graham.github.io/paypulse-paypal-ai-agent/

Built With

Share this project:

Updates

Submission history