Karnataka unveils Tathyakosh, putting 1.5 million datasets on one searchable platform
Karnataka has launched Tathyakosh, a unified data discovery platform bringing together more than 1.5 million datasets from 458 sources.
India has plenty of public data. The harder part is finding it.
Researchers, startups and developers often have to look through multiple government websites, research portals and institutional repositories before they can identify a dataset that fits their needs. A new platform from Karnataka is attempting to make that process simpler.
The Government of Karnataka’s Centre of Excellence in Artificial Intelligence, ACSEL Technology Forum, has launched Tathyakosh, a platform that brings information about more than 1.5 million datasets together in one searchable interface. The datasets have been sourced from 458 repositories.
Tathyakosh was launched in Bengaluru on August 12 by Karnataka IT-BT Minister Priyank Kharge at Vikas Soudha. The platform is intended for researchers, policymakers, startups, developers and data scientists. At the launch, Kharge highlighted the importance of making India's existing data easier to access as the country builds its AI capabilities.
“India’s AI ambitions will depend not only on talent and computing power, but also on our ability to find and use the vast amount of data that already exists across our institutions.”
He added that Tathyakosh is being built as an open, non-commercial Digital Public Good to make data more discoverable and easier to use for people working with Indian data.
A single place to explore public data
The idea behind Tathyakosh is not to create another database where users have to upload or store information. Instead, it works as a discovery layer that helps people locate datasets available across different sources.
The platform currently covers 16 areas, ranging from health and agriculture to education, infrastructure, environment, economy and governance. Its index includes 346 Indian sources alongside 37 international repositories that contain data relevant to India.
Users can narrow down datasets using details such as location, subject area, time period, file format and access type. According to the launch information, more than 92% of the indexed datasets are open access and 95% are machine-readable, which means they can be more readily used by software and computational tools.
Tathyakosh also attempts to answer another question that comes after finding a dataset: can it actually be used?
Each entry receives an assessment based on seven factors, including accessibility, freshness, completeness, granularity and machine-readability. This is intended to give users a quick indication of the usefulness and quality of the available data. The platform’s historical coverage currently extends back to 1947.
Building an easier data layer for AI
The launch comes as demand for usable datasets grows across India’s AI ecosystem. Building AI applications often requires large volumes of structured information, but developers first need to know what data exists and whether it is suitable for their particular application.
Tathyakosh could help reduce some of that initial search work, particularly for projects involving healthcare, agriculture, climate, legal research and financial inclusion. Universities and research teams can also use the platform to explore data across different sectors.
Karnataka has a specific focus on the platform, with more than 50,000 datasets related to the state already indexed. These include sources such as the Karnataka Open Government Data Portal, Mahiti Kanaja, KSRSAC/K-GIS, the Indian Institute of Science and the Raman Research Institute.
Dr. N. Manjula, Secretary, Department of Electronics, IT and BT, Government of Karnataka, said: “Tathyakosh gives users a way to find that information without having to search each source separately.”
She also pointed to the platform’s ability to help users assess whether datasets are reliable and suitable for their intended use, potentially saving time for researchers, developers and data scientists.
Tathyakosh has been designed as an open, non-commercial Digital Public Good. It does not take ownership of the datasets or store them on its platform. Instead, users are directed to the original source when they want to access the underlying data.
As more government and research datasets become available, the platform is expected to expand its coverage. For India’s growing AI and data ecosystem, its larger role could be helping users discover existing information before spending time and resources looking for it elsewhere.
Tathyakosh is now live at Tathyakosh.in.

