BengkhelSoftware & Data Studio
ID Contact us

Software & data studio from Indonesia

Websites, web apps, and dashboards with numbers you can trust.

We're a two-person team, a software engineer and a BI analyst. We build systems that are easy to use and replace manual spreadsheet reporting.

Tell us about your project See our work

Work

Web appMobile-firstLive

Karyanusa

An operations and finance system for a rattan export business: from a single purchase order to a net-profit figure the owner can trust.

Visit site
Sign in
Karyanusa sign-in screen with a rattan background
Home
Owner home: estimated profit, incoming revenue, and outstanding obligations
Money
Money screen: profit per order and business profit per month
New order
New order form with products and buyer payment terms

Problem

Orders, subcontractor work orders, and cash flow lived on paper and in spreadsheets. Real profit was unknown, and money leaking to subcontractors only surfaced after a loss.

Solution

One mobile-first app with three roles (Owner, Sales, Office). Each role sees only the numbers it should, and those rules are enforced in the database itself.

Outcome

  • 1 net-profit numberper order and per month
  • 3 roleswith separate access rights
  • Subcontractor advancestracked automatically
  • Next.js
  • TypeScript
  • Tailwind CSS
  • Supabase
  • PostgreSQL
  • Vitest
  • Vercel
Technical details

Key features

  • Role-based dashboards. Owners see profit and cash flow, Sales sees order value and receivables, Office sees production progress with no money figures.
  • Subcontractor work orders. Split an order across subcontractors, track goods received and returns, and flag unreturned advances.
  • Multiple currencies. The exchange rate is locked per order at signing, so old orders don't drift with the market.
  • Final profit. Closing an order locks its profit based on actual costs.

Engineering decisions

  • Money security in the database. Access is enforced with Postgres Row Level Security. Costs a role can't read come back as null, not 0, so profit is never shown wrong.
  • One source for every number. No stored totals; all calculations live in one pure module that's tested without a database.
  • Void, never delete. Money rows are never edited or deleted, so there's always an audit trail.

How we build

Built with AI, guarded like an engineering team

We use an AI coding agent to move at the speed of a much larger team, without dropping the safeguards professional engineering teams rely on.

  1. An engineer steers

    One engineer acts as product owner and architect. The AI coding agent writes code within clear direction and constraints.

  2. Everything goes through PRs

    Every change is checked by an AI reviewer, then must pass human review before it reaches the main branch.

  3. Tested automatically

    Karyanusa runs 450+ unit tests plus contract tests against a real database on every change.

  4. Critical rules in the database

    Access policies and schema changes get their own review, because mistakes there don't always fail a test.

Services

Two skill sets, one team

When a project needs both, we build it together with no brief handed back and forth.

Websites & web apps

Led by Adiyansa, Software Engineer

  • Company websites & landing pagesFast, mobile-friendly, and measurable.
  • Online storesCatalog, cart, payments, and shipping.
  • Web apps & internal systemsOrders, production, and finance in one system.

Data & Business Intelligence

Led by Farhan, Business Intelligence Analyst

  • BI dashboardsDatabases, spreadsheets, and apps combined in one dashboard.
  • ForecastingDemand prediction for inventory and planning.
  • Report automationReplacing days of manual work.

Team

The people behind Bengkhel

Photo of Adiyansa Wicaksana

Adiyansa Wicaksana

Software Engineer

LinkedIn
  • 99% payout successacross IDR 5.4 trillion in transactions
  • Anti-fraud & AMLbuilt and integrated end to end

In fintech, logistics, and digital agencies since 2022: Stockbit, Paper.id, Logisly, Suitmedia, Pintarnya. B.Sc. Computer Science, ITB.

Photo of Farhan Imam Naufal

Farhan Imam Naufal

Business Intelligence Analyst

LinkedIn
  • Forecast error from 12% to 4%automated forecasting, 2 days of work down to 1 hour
  • 3 days to 15 minutesautomated processing of 600K+ rows of data

In logistics, travel, and edtech: SiCepat, tiket.com, Schoters by Ruangguru. B.Sc. Statistics, Universitas Diponegoro; Bank Indonesia scholar.

Tell us about your project

A rough outline is enough. We reply within 1 business day with questions or a time to talk.

Emailfarhan@bengkhel.com WhatsApp+62 851 5614 3648