OVERVIEW

ERPNext is an open-source ERP built on the Frappe framework. It's widely used in manufacturing, especially for make-to-order and make-to-stock operations. This skill is tailored for furniture manufacturing — covering BOMs, production routing, timber/hardware inventory, and quality inspection.

ERPNext exposes a comprehensive REST API for every doctype, plus an RPC API for server-side methods. This skill gives AI agents the knowledge to query, automate, and report on a furniture factory — without hallucinating endpoints or field names.

ARCHITECTURE

The recommended pattern uses a Cloudflare Worker as a proxy — this avoids CORS, secures API keys at the edge, and enables AI analysis via Workers AI.

Base URL

GET  https://<site>/api/resource/<DocType>
POST https://<site>/api/method/<dotted.path>

Authentication

# API Key + Secret (recommended)
Authorization: token api_key:api_secret

# Or Basic Auth
Authorization: Basic base64(user:password)

CAPABILITIES

What your agent can do with this skill installed:

• Manage BOMs — create, explode, and cost Bills of Materials for furniture
• Schedule production — work orders, job cards, workstation capacity
• Track inventory — stock levels, reorder alerts, material transfers
• Process sales — orders, deliveries, invoices, customer analytics
• Handle procurement — purchase orders, supplier performance, cost tracking
• Quality control — inspections, readings, pass/fail automation
• Generate reports — revenue, production output, material usage, margins
• Natural language queries — translate questions into ERPNext API calls
• Deploy to edge — Cloudflare Workers proxy with AI analysis

KEY DOCTYPES

ERPNext uses "doctypes" — each is a data model with its own REST endpoint.

Item Groups (furniture setup)

MANUFACTURING FLOW

Operations Routing (typical furniture)

Example BOM — Dining Table

{
  "item": "NOHA-DINING-TABLE-8",
  "quantity": 1,
  "items": [
    {"item_code": "RM-OAK-25MM",     "qty": 24,  "uom": "Board Foot"},
    {"item_code": "RM-OAK-VENEER",   "qty": 2.4, "uom": "Square Meter"},
    {"item_code": "HW-BOLT-M8X60",   "qty": 16,  "uom": "Nos"},
    {"item_code": "HW-BRACKET-L",    "qty": 8,   "uom": "Nos"},
    {"item_code": "FN-DANISH-OIL",   "qty": 0.8, "uom": "Litre"},
    {"item_code": "FN-SANDPAPER-180","qty": 4,   "uom": "Sheet"}
  ]
}

API PATTERNS

List Records

GET /api/resource/Item
  ?filters=[["item_group","=","Finished Goods"]]
  &fields=["name","item_name","standard_rate","stock_uom"]
  &limit_page_length=50
  &order_by=item_name asc

Active Work Orders

GET /api/resource/Work Order
  ?filters=[["status","in",["Not Started","In Process"]],["docstatus","=",1]]
  &fields=["name","production_item","qty","produced_qty","status","planned_start_date"]
  &order_by=planned_start_date asc

Explode BOM (get all materials for a batch)

POST /api/method/erpnext.manufacturing.doctype.bom.bom.get_bom_items
{"bom":"BOM-NOHA-DINING-TABLE-001","qty":30,"fetch_exploded":1}

Stock Balance

POST /api/method/erpnext.stock.utils.get_stock_balance
{"item_code":"RM-OAK-25MM","warehouse":"Raw Materials Store"}

Filter Operators

REPORTING PATTERNS

ERPNext supports server-side aggregation via frappe.client.get_list with group_by.

Revenue by Customer

POST /api/method/frappe.client.get_list
{
  "doctype": "Sales Invoice",
  "filters": [["docstatus","=",1],["posting_date",">=","2026-01-01"]],
  "fields": ["customer","sum(grand_total) as revenue"],
  "group_by": "customer",
  "order_by": "revenue desc",
  "limit_page_length": 20
}

Production Output by Product

POST /api/method/frappe.client.get_list
{
  "doctype": "Work Order",
  "filters": [["status","=","Completed"],["actual_end_date",">=","2026-03-01"]],
  "fields": ["production_item","sum(produced_qty) as total_produced"],
  "group_by": "production_item",
  "order_by": "total_produced desc"
}

Material Usage This Month

GET /api/resource/Stock Ledger Entry
  ?filters=[["voucher_type","=","Stock Entry"],["posting_date",">=","2026-03-01"],["actual_qty","<",0]]
  &fields=["item_code","sum(actual_qty) as total_issued"]
  &group_by=item_code
  &order_by=total_issued asc

AI INTEGRATION

Natural language in, operational intelligence out.

Cloudflare Worker Proxy

async function proxyERPNext(request, env) {
  const url = new URL(request.url);
  const erpPath = url.pathname.replace('/api/erp/', '/api/');
  const erpUrl = `${env.ERPNEXT_URL}${erpPath}${url.search}`;

  return fetch(erpUrl, {
    method: request.method,
    headers: {
      'Content-Type': 'application/json',
      'Authorization': `token ${env.ERPNEXT_API_KEY}:${env.ERPNEXT_API_SECRET}`,
    },
    body: request.method !== 'GET' ? await request.text() : undefined,
  });
}

System Prompt for Query Generation

You are an ERPNext API expert for a furniture manufacturer.
Generate the API call for the user's question.

API: GET /api/resource/<DocType>?filters=...&fields=...
RPC: POST /api/method/<dotted.path>

Key doctypes: Item, BOM, Work Order, Job Card, Stock Entry,
Sales Order, Purchase Order, Quality Inspection, Bin.

Filters: JSON arrays [field, operator, value].
Always include docstatus=1 for submitted docs.
Respond with HTTP method, path, and body only.

INSTALL

Works with Claude Code, Cursor, Windsurf, Cline, and 30+ other AI agents:

npx skills add imbilawork/2nth-skills@erpnext-furniture

Or install globally with no prompts:

npx skills add imbilawork/2nth-skills@erpnext-furniture -g -y

The skill loads automatically when your agent works on ERPNext or manufacturing tasks.

GOTCHAS