Banya LabsBanya Labs/Apps/Leriso
In progress

App 07  ·  FIG. 007

Leriso logo

Leriso

Take Orders Fast. Track Orders Live.

POSResturant
App not yet live

Customer Target

Southern African Fast-Casual Restaurants, Bars & Bistros

Status

In progress

Agent Surface B

MCP Ready

Model Context Protocol Enabled

Registered in /.well-known/mcp.json & /llms.txt with deterministic Zod schema validation.

About Product

### Speed-Optimized Point of Sale KafuPOS is a high-performance POS system designed specifically to handle the high-turnover requirements of fast-casual dining environments in Southern Africa. The application natively supports a multi-currency environment (ZAR, USD, etc.) with offline-tolerant resilience.

### Key Architectural Pillars 1. **Cashier Register (0.8s Touch Grid):** Instant item selection, flexible order customization, and immediate ticket dispatch. 2. **Live Kitchen Monitor:** An interactive, sound-gated display showing cooking queues, ticking ticket times, and single-tap order status completions. 3. **Back-Office Administration:** Centralized workspace to configure menus, custom item modifiers, currency rates, and review comprehensive ledgers. 4. **Strict Audit Trail:** Tracks all operational actions (Login, Mode Switch, Order Actions, Menu Edits) for complete transparency and leakage prevention.

### Technical Stack - **Framework:** Next.js (App Router) & React 19 - **State Management:** Zustand (lightweight client-state synchronization) - **Styling & UI:** Tailwind CSS, Lucide Icons, and Base UI

Dual-Surface Execution

Human vs. Agent Flow

Every action in Leriso is mapped bi-directionally across Surface A (Visual UI) and Surface B (Programmatic MCP Engine).

Surface A — Human Ergonomics Flow
  1. 01Cashiers, kitchen staff, and admins input their secure PINs to access their roles on the shared POS terminal.
  2. 02The cashier opens the Cashier workspace, selects items on the 0.8s touch grid, configures modifiers, and submits the order.
  3. 03The order is instantly dispatched to the kitchen monitor, sounding an alert and displaying active tickets under preparation.
  4. 04Kitchen staff mark the order ready, notifying the cashier who handles customer payout and dispatches the food.
  5. 05Administrators access the Admin workspace to view audit logs, adjust menu lists, modify pricing, and configure currencies.
Surface B — Agentic Machine Execution Rails
Protocol: Model Context Protocol (MCP 1.0)Auth: Bearer banya_ag_...

1. Discovery: Scraper/Agent fetches /llms.txt or /.well-known/mcp.json manifest.

2. Tool Invocation: Executes deterministic Zod-validated tool contracts matching leriso operations.

3. Replay Protection: Enforces Idempotency-Key: <uuid> to prevent duplicate mutations during LLM retry loops.

4. Token Compression: Returns structured, PII-minimized Markdown/JSON stripping all layout noise.