Agentic Process Engineering (APE)
Do not start with agents. Start with process.
APE: designing AI agent systems as executable business processes — tools, roles, knowledge & procedures, controls & metrics. For enterprise AI integration.
casys.ai · APE canvas · LinkedIn · hello@casys.ai · Taiwan
| Platform | Status | Scope |
|---|---|---|
| eInvoice Platform | Live · French mandate 2026 | E-invoicing platform. Process-facing invoice lifecycle tools + provider adapters. Iopole live & self-serve. Other adapters onboarded with support. |
| ERP Platform | Live beta | MCP-driven ERP workflows. ERPNext is the most mature adapter. Dolibarr mapping underway. |
Public. MIT.
| Repository | Scope | Signals |
|---|---|---|
| mcp-server | Core MCP server |   |
| mcp-einvoice | MCP server for French e-invoicing. PA-agnostic adapter pattern. |   |
| mcp-erpnext | MCP server for ERPNext / Frappe ERP. Interactive UI viewers. |   |
| Repository | Scope |
|---|---|
| mcp-erp | ERP-agnostic MCP adapter layer: ERPNext, Dolibarr, more.  |
| mcp-geomatic | Private. MCP server for geospatial operations: QGIS Server + PostGIS + GEOS (WASM), natural-language geo data ops. |
| Repository | Scope |
|---|---|
| DenoClaw | Deno-native AI agent framework: Subhosting + Sandbox + A2A, zero Node.js deps.  |
Process artifacts, source code, governed action.