irwan.world

World hub · data & technology

irwan.world

Connecting local languages and heritage to the world.

A data engine for endangered languages and manuscript digitization — built in Sumatra, for the world.

Services

What irwan.world offers

Five concrete capabilities, one pipeline — from raw archives on a shelf to machine-readable language data, and the infrastructure to serve it.

Language datasets

Corpus building from books and archives — segmentation, labeling, licensing, provenance.

OCR & digitization

Photo-to-searchable-text pipelines for regional-language documents.

Tourist ↔ local translation

Text and speech translation built on top of the Li Niha dataset.

Manuscript restoration

Page-by-page restoration and print preparation of heritage scriptures.

Data infrastructure

Self-hosted inference, OCR API, and encrypted tunnels — private by default.

Project 01

Li Niha — the Nias Language Dataset

Nias (Li Niha) is spoken by ±1 million people on Nias Island, North Sumatra — with near-zero NLP support. This dataset changes that: a clean, licensed, provenance-tracked foundation for teaching machines to read and translate a language the industry forgot.

71,339 corpus words · 30 source books
41,971 labeled word images
7,807 unique words
59 characters · incl. õ ö ô ’
98.8% character accuracy · OCR baseline
97.7% word accuracy · OCR baseline

CRNN + CTC baseline, trained end-to-end on a consumer CPU — no GPU.

How the dataset is built

01

Acquire & digitize

30 source books from the “Li Nono Niha” series, photographed page by page.

02

Segment & label

41,971 labeled word images across a 59-character alphabet, incl. õ ö ô ’.

03

Train & measure

CRNN + CTC baseline trained on a consumer CPU — 98.8% character, 97.7% word accuracy.

04

Publish & track

71,339-word corpus, 7,807 unique words — every example records its book and page.

Built on top of the dataset

Tourist ↔ Local Translator

For tourism and surfing destinations across Nias — a bridge between visitors and the local community: text and speech in, text and speech out.

T2T · STT · TTS · S2S

OCR & Document Digitization

Turn Nias teaching books, church and government archives into searchable digital text — from shelf photo to corpus.

OCR · DIGITIZATION

Project 02

Project-15 — 《佛說佛名經》

Buddha Names Sūtra · restoration & print preparation

Restoration and print preparation of the Buddhist scripture 佛說佛名經, Volume I & II — 28 Katern per volume, CorelDRAW sources, 380 × 260 mm print sheets.

Red Hanzi in vertical columns, blue pinyin at their side — the visual grammar of the scripture, preserved exactly.

Working standards

Source integrity

Every file is SHA-256-hashed before and after work — sources are never mutated.

Isolated runs

Every job happens in its own separate, timestamped run.

Reopen audit

Every page is saved, closed, and reopened before it is counted as done.

Owner QC

Every page is personally approved by the owner before the next one begins.

Proof of engineering — 4-engine OCR benchmark

HunyuanOCR
100%

100% Hanzi accuracy

GLM-OCR
50%

FAIL — characters got duplicated

Key finding: reading the vector text directly from the source file is 100% accurate — outperforming every raster OCR engine.

Deliverables

Restored page files Print-ready 380 × 260 mm sheets SHA-256 integrity manifest
Volume I — final audit Volume II — read-only preparation

Infrastructure

From homelab to the world

  • Self-hosted GPU inference node (48 GB VRAM) running local OCR & LLM models.

  • OCR API (PaddleOCR PP-OCRv6) on a Linux server — localhost only, reached through an encrypted tunnel, never publicly exposed.

  • HunyuanOCR handles the Project-15 restoration line — 100% Hanzi accuracy, best of four OCR engines benchmarked.

  • An encrypted tunnel bridges the private network to the internet — irwan.world DNS is the public doorway.

Request flow

public irwan.world
edge / tunnel encrypted
origin homelab & VPS
services OCR API · models · datasets

Our Network

Four sites, one ecosystem.

irwan.world is part of a four-site ecosystem — data & technology, print, network infrastructure, and AI.

Percetakan Bintang

Trusted print shop in Medan — invitations, banners, business cards, books, and other print needs.

Visit →
Bintang Network

Network infrastructure, e-voting, cloud ECS/EDS, plus web design, hosting, and domains.

Visit →
EntityHub

AI studio — AI training, model fine-tuning, and web/desktop/mobile app development.

Visit →

Contact

Let’s talk data, language, or heritage.

location Indonesia