Sri Lanka has accumulated an enormous amount of public information over decades. Laws, court judgments, gazettes, company records, economic statistics, tax regulations, government circulars, parliamentary documents and regulatory publications are available across hundreds of institutions and websites. The problem has often not been the absence of information, but the difficulty of finding, connecting, updating and using it.
Lanka Data Net (LDN) is attempting to address that problem by building an infrastructure layer for Sri Lankan public information. At the centre of that infrastructure is OpenPIN, which Lanka Data describes as its connectivity and data-harvesting layer.
What is OpenPIN?
OpenPIN can be understood as the connection layer between Sri Lanka’s information sources and the Lanka Data platform.
Instead of requiring every application to independently connect to government websites, registries, document repositories and other information sources, OpenPIN is intended to provide a common mechanism for acquiring and processing information. Lanka Data describes mechanisms including APIs, document acquisition, web connectors, structured feeds, scheduled collection, incremental updates and change detection.
The basic concept is:
Public information sources → OpenPIN → Lanka Data → AI / APIs / applications
This architecture is important because the same underlying information can then be reused by different services rather than being collected repeatedly by each application.
From Documents to Structured Data
Lanka Data Net says it systematically collects, digitises, organises and indexes information covering areas such as legislation, judicial decisions, gazettes, parliamentary material, Central Bank statistics, taxation, business information, regulatory publications and other public datasets.
The objective is not simply to store PDFs.
The information is processed and standardised so that it can become searchable, structured and machine-readable.
This creates an important progression:
Documents → structured data → indexed information → connected knowledge → AI
That foundation supports the various services developed around Lanka Data Net.
The Lanka Data Net Architecture
Lanka Data describes its broader architecture as a combination of a data foundation and an intelligent layer.
At the data level, it maintains curated and continuously updated datasets across areas including law, taxation, business, news, pricing and education. Those datasets are standardised and indexed.
The next layer is the RAG — Retrieval-Augmented Generation — infrastructure.
When a user asks a question, the system can identify the relevant domain, retrieve information from the underlying repositories and use an appropriate AI model to produce the response. Lanka Data describes this as dynamic domain routing to specialised AI experts.
The architecture can therefore be represented as:
The Data Foundation
Lanka Data currently lists repositories covering a wide range of public information.
These include:
- Supreme Court judgments
- Court of Appeal judgments
- Acts and amendments
- Bills
- Gazettes
- Parliamentary material
- Central Bank publications
- economic indicators
- taxation
- registered companies
- listed companies
- CSE announcements
- intellectual property
- customs and trade
- medicines
- government circulars
- government forms
- public-safety information
- disaster and weather information.
Lanka Data says its information is sourced from official institutions and that the original official content is preserved rather than altered.
It currently advertises more than 90,000 public documents, alongside much larger figures for its structured data repositories. It also advertises an API platform with more than 100 endpoints. These figures are published by Lanka Data itself and should therefore be understood as platform claims rather than independently audited measurements.
Why OpenPIN Matters
The significance of OpenPIN is that it addresses a problem that becomes increasingly difficult as the number of data sources grows.
Consider a legal AI platform.
It might need to collect:
- Acts
- Amendments
- Gazettes
- Bills
- Court judgments
- Law Reports
- circulars
- regulations.
A tax platform requires a different set of sources.
A business intelligence platform requires company records, financial information and economic statistics.
If each application builds its own collection system, there is considerable duplication.
OpenPIN provides a common acquisition architecture so that information can potentially be collected once and reused across multiple applications.
That is the underlying infrastructure idea.
OpenPIN and Chat2Find
OpenPIN is not itself the chatbot.
The relationship is better understood as:
OpenPIN
collects and connects information.
Lanka Data Net
organises and maintains the structured data repositories.
RAG/domain experts
retrieve relevant information.
Chat2Find
provides language understanding, reasoning and the conversational interface.
Lanka Data explicitly describes the combined system as a data-to-intelligence architecture, where information is collected, structured, indexed, retrieved, interpreted and finally delivered through AI.
This distinction is important.
The AI model is not intended to be the sole source of knowledge. The underlying Lanka Data repositories provide the information that the AI retrieves at query time.
The API Layer
Another important component is the Lanka Data API.
The platform currently advertises more than 100 API endpoints and provides structured JSON access to information such as company records, court judgments, stock information, pharmaceutical registrations and intellectual-property records.
This means Lanka Data is moving beyond a conventional website.
Developers can potentially use the information as an infrastructure service inside their own applications.
A recent example is Daily FC’s integration of the Lanka Data API into its editorial platform to access structured public information for journalism and research.
The AI Layer
The intelligent layer sits above the structured data.
Lanka Data says its system includes:
- Lanka Data Search
- domain-specific RAG models
- Chat2Find base models
- fine-tuned Sinhala, Tamil and English capabilities.
Lanka Data also reports the release of a 255-million-plus-token trilingual conversational corpus, containing approximately 279,248 records in Sinhala, Tamil and English, including code-mixed forms such as Singlish and Tanglish.
This creates a broader architecture in which local data and local-language AI are developed together.
From LankaLaw to LankaBIZ
The infrastructure is designed to support multiple applications rather than one AI product.
Three prominent examples are:
LankaLAW — legal information and research
LankaTAX — tax information and assistance
LankaBIZ — business and economic intelligence
Lanka Data describes these as applications built on a common underlying data infrastructure rather than completely independent information systems.
This creates an important platform effect:
One data infrastructure → multiple specialised applications.
The Larger Vision
The significance of OpenPIN and Lanka Data Net therefore extends beyond another search engine or chatbot.
The larger proposition is the creation of a Sri Lankan public-information infrastructure in which fragmented information can be continuously collected, structured, indexed and made available to both humans and machines.
The architecture can ultimately be viewed as:
The central idea is therefore not simply “build an AI chatbot.”
It is:
Build the data infrastructure first, then build intelligence on top of it.
That is what makes OpenPIN and Lanka Data Net strategically significant. Their long-term value will depend not only on the amount of information they collect, but on the quality, provenance, freshness, structure, connectivity and accessibility of that information.
And that is ultimately where the competition in Sri Lankan data infrastructure is likely to be decided: not by who has the largest collection of documents, but by who can turn fragmented national information into reliable, connected and machine-readable knowledge.
