Gyanmandir Library - Sanskrit Book Database
About Gyanmandir Library
Our client manages a vast database of over 2 million records of Sanskrit books and approached MagnusMinds with a critical performance issue impacting their application's search functionality. The inefficiency in handling such a large dataset and the inability to accommodate user input errors in search queries had led to a frustrating user experience.

MagnusMinds transformed the Gyanmandir Library's search experience by replacing SQL Server full-text search with Elasticsearch. The result was a 10x speed improvement and a dramatically more forgiving search interface — users can now find Sanskrit books even with spelling errors or partial titles, making the platform genuinely usable for scholars and researchers.
Our Approach
Challenge & Solution
The Challenge
The client's application struggled with performance due to the size of the dataset — more than 2 million book records, all written in Sanskrit. The SQL Server backend was unable to efficiently manage search queries at this scale, resulting in significant delays and long wait times before results appeared. A second major challenge was search accuracy: users frequently entered incorrect or incomplete book titles, but the existing search lacked tolerance for minor errors such as spelling mistakes or partial inputs. Searches failed when users did not provide an exact match, causing frustration when they couldn't retrieve relevant books despite minor input errors.
Our Solution
To resolve the performance bottleneck, Elasticsearch was integrated as the primary search engine, replacing the traditional SQL search approach. Searches that once took an excessive amount of time were now carried out up to 10 times faster, providing users with a far more responsive experience. To address inaccurate and incomplete queries, Elasticsearch's advanced fuzzy search capabilities were leveraged — enabling the application to retrieve relevant results even when the user's input was not an exact match, accommodating spelling mistakes and partial titles with ease.
What We Built
Key Features
Elasticsearch integration replacing SQL Server search
Fuzzy search for Sanskrit text with error tolerance
Partial title matching for incomplete search queries
10x faster search across 2M+ book records
.NET Core 8 backend for modern performance
Scalable architecture for growing library collections
Impact
Results & Outcomes
Search performance improved 10x over SQL Server
Fuzzy and partial search for 2M+ Sanskrit book records
Near-instantaneous results replacing long wait times
Users no longer hindered by minor spelling mistakes
Stack
Technologies Used
Client
Gyanmandir Library
Industry
Digital Library / Sanskrit Books
Technologies
Elasticsearch, .NET Core 8
Have a Similar Project?
Let's discuss your requirements and build something extraordinary together.
