Introduction
SharpOMatic is an open-source workflow builder focused on AI-driven tasks designed for .NET projects.
Configuration over code
One of the biggest challenges when building an AI-focused workflow is cycle time for trying ideas. AI is notorious for needing lots of experiments to find the right combination of model, prompts, tool calls, and glue logic to get the outcome you need. Using SharpOMatic, you can prefer configuration over code, which lets you iterate on ideas much faster.
Evaluations
All AI based workflows should have evaluations. They compound the benefit of configuration over code because your evaluation is run for each new prompt, design change and model release. Objective measures are needed for reliable production systems, not just eye balling results.
Host your own execution
The execution engine is hosted in your project so you retain complete control over the environment and storage of all data. You can run with OpenAI, Azure OpenAI, and Google connectors while still keeping all workflow and run data in your own system. This makes it easy to merge workflows into an existing backend because it really is part of the that same backend.
Deep integration
SharpOMatic is a .NET native project, so you can use familiar C# for glue logic in workflow code nodes. You can call from code nodes directly into your backend code for fast, simple integration. Expose your C# types for structured outputs and C# functions for tool calling.
What you can do
- Build workflows with nodes for models, code, branching, and orchestration.
- Run and debug workflows with rich run state and persisted history.
- Configure and run evaluations against datasets with grader workflows.
- Embed the editor into your own ASP.NET Core host.
- Call into your existing backend code from the workflow.
Project layout
- Frontend UI:
src/SharpOMatic.FrontEnd - Editor host:
src/SharpOMatic.Editor - Workflow engine:
src/SharpOMatic.Engine - Sqlite for storage:
src/SharpOMatic.Engine.Sqlite - SQL Server for storage:
src/SharpOMatic.Engine.SqlServer - Sample host:
src/SharpOMatic.DemoServer
What next?
Use the Getting Started guides to get up and running in just a few minutes.
Read the Core Concepts section before continuing to more specific features.
To evaluate your workflows, go to Evaluations.