A trustworthy, federated architecture for turning fragmented data into AI-ready intelligence — without centralising it.
Data Unlock is a Digital Public Infrastructure (DPI) framework designed to help governments, public institutions, and their technology partners convert siloed, human-readable datasets into interoperable, machine-interpretable, and AI-ready data products. It does this while preserving data ownership, enforcing governance, and enabling purpose-bound access.
| Audience | What you'll find here |
|---|---|
| Government agencies | A framework for making your data discoverable, linkable, and usable across departments — without surrendering control. |
| IT vendors & system integrators | Implementable technical specifications, protocol choices, and reference architectures you can build against. |
| DPI architects & policymakers | Design principles, business architecture, and a compliance checklist for deploying Data Unlock as national or state-level infrastructure. |
| Researchers & data scientists | Standards for accessing AI-ready government data through APIs, MCP servers, knowledge graphs, and vector stores. |
Most governments already have the data. What they lack is a mechanism to make that data trusted, linked, machine-readable, and accessible — in a way that respects institutional boundaries. Data Unlock provides that mechanism through a four-stage value chain:
- Prepare — Standardise, certify, and catalogue raw data into trusted data products.
- Connect — Resolve entities, align semantics, and enable federated queries across sources.
- Enable — Transform data into AI-ready representations: embeddings, knowledge graphs, vector stores.
- Apply — Deliver real-world outcomes through dashboards, conversational AI, and workflow automation.
Each stage is backed by open standards, open protocols, and (where available) open-source reference implementations.
-
Data Boarding Pass — A declarative manifest that describes a specific Data Unlock deployment: who uses it, what data flows through it, for what purpose, and through what interfaces. Learn more →
-
Data Passport — The machine-readable metadata envelope that travels with a data asset as it moves through the ecosystem — combining provenance credentials, access policies, and structural metadata. Learn more →
- Start with the Overview to understand the concept.
- Read the Architecture section to understand roles and technical structure.
- Dive into the Technical Specifications for implementable protocol details.
- Follow the Implementation Guide to plan your deployment.
- Study the Reference Implementations to see how India is applying this framework.
Data Unlock is an initiative of People+AI, an EkStep Foundation initiative. This specification is published as an open framework for adoption by any government, institution, or technology partner worldwide.