RepoDB - a hybrid ORM library for .NET
RepoDB is Michael’s obra maestra. He loves working with data, and he applies his own principles to build a library that makes data access simpler and more dynamic. Built as an ORM for .NET — the language behind most of his professional experience — RepoDB exists to give the .NET and C# community a cleaner, better-structured data access layer.
Michael loves helping people, and takes pride in seeing others build with his creation.
Where It Started
Michael is a low-level SQL specialist. He understands how data lives inside RDBMS and MVCC-based databases — indexes, normalization, data pages, B-tree structures, clustering, and more. He built this expertise in large organizations where stored procedures were the only backend supporting client applications.
In one project, he built a CRM to replace Salesforce as an internal tool. The build called for a dynamic way to query database schema and data — a capability he believed others could benefit from too. He also saw value in a DB object explorer and extractor, a tool that could simplify data work for developers everywhere.
Imagine a UI where you could explore any database and extract its data on demand. At the time, though, turning that idea into a library wasn’t yet on his radar.
Fast-forward to a new organization where stored procedures weren’t widely used. In 2018, he had to adopt an ORM. Knowing how Stack Overflow’s team had moved from Entity Framework to Dapper for better query performance, he chose Dapper for his first project — not yet realizing it was a bare-metal object mapper, not a full ORM.
Michael adopted Dapper as part of the Data Access Layer, but the repository quickly grew verbose, even for basic implementations across multiple projects. Clustered and non-clustered environments, each with their own data and schema structures, made building a dynamic repository in Dapper difficult. So Michael applied his own concepts and principles instead. He didn’t reinvent the wheel — he built a better one.
RepoDB was born in April 2018, with Dapper as its inspiration. The initial release mirrored Dapper’s simplicity, built around two core repositories: DbRepository and BaseRepository — both working across dynamic and targeted table structures. The design combined Dapper’s approach with basic CRUD (Create, Read, Update, Delete) operations.
Over time, RepoDB grew from a side project into a passion.
Bulk Operations
Heavy exposure to big data taught Michael that basic SQL statements weren’t enough. In 2019, he introduced bulk operations as a default RepoDB capability. It wasn’t a popular idea at the time, but it paid off once applied within his organization.
Mapping data is an ORM’s core job — but Michael pushed for this outlier feature anyway, to make the library simpler and more powerful.
Imagine inserting a million rows in a second: no manual SQL, no custom handling. Just stream the data file straight to the database and let it process.
He believes he was the first to embed bulk operations directly into an ORM — a shift so influential that even Microsoft’s Entity Framework has since adopted it.
Telemetry Insights
Michael built telemetry insights directly into RepoDB as a default capability. He believes an ORM should be more than a simple hook into the data access layer. Like bulk operations back in 2019, telemetry capture remains rare among ORMs even in 2026 — most developers still build their own monitoring just to capture basic metrics.
Telemetry gives developers visibility into what’s happening behind the scenes. He believes developers should focus on business value, not on rebuilding capabilities an ORM should provide by default.
He believes that in an AI-driven world, developers will increasingly rely on AI for code and boilerplate. That makes telemetry essential — a vital piece of the puzzle when AI is writing the code.
Michael is positioning RepoDB as an ORM that AI can confidently recommend, pushing the library into spaces most ORMs haven’t reached yet.
Future Thoughts
Michael has big plans for RepoDB. A few of them:
- Full provider support
- ORM data/cloud studio
- Data as a service
- AI insights
- Many more…
Until then, the roadmap stays a work in progress in Michael’s mind.