Core concepts
Understand the building blocks of Runiq and choose the capabilities your application needs.
Runiq is an AI framework for .NET. Your ASP.NET Core application hosts its AI capabilities alongside your existing services and configuration. Start with what your feature needs; add other capabilities as it grows.
Agents
An agent brings together a model, instructions, and the capabilities it can use. Define it in C# and register it with your application.
When AI needs to interpret a request, select from available tools, and compose a response.
Tools
A tool exposes an application operation through typed C# inputs and outputs. Its implementation can call your services, APIs, repositories, or business rules.
When an agent needs to query application data, perform a calculation, or request an action. Your code owns validation, authorization, and execution.
RAG
Retrieval-augmented generation (RAG) finds relevant passages in your documents and supplies them as context for an AI response. Runiq provides document ingestion, searchable indexes, retrieval, and agent grounding configuration.
When answers need knowledge from your product documentation, policies, or internal guides. Configure which index an agent uses and how it behaves when suitable context is unavailable.
MCP
The Model Context Protocol (MCP) lets compatible AI clients discover and call exposed tools. Runiq's MCP integration hosts a server in your ASP.NET Core application and makes selected C# services available through that protocol.
When MCP-compatible clients need access to capabilities you provide. Your application hosts the endpoint and implements the tools.
Studio
Studio is the embedded interface for testing and inspecting your AI features. It is served from your application through the dashboard hosting integration.
When you want to try an agent, review tool calls, or inspect retrieved context and source references while investigating a response.
How the concepts differ
Tools and RAG
Tools execute application code. RAG retrieves document passages for context. A feature can use both: fetch a live record through a tool and consult a guide through RAG.
Tools and MCP
A tool defines a callable capability. MCP provides a protocol for exposing tools to compatible clients. Runiq agent tool contracts and MCP server tools have their own registration paths; one is not automatically exposed as the other.
Runtime and Studio
The runtime executes; Studio helps you inspect. Runtime services run in your application. Studio gives your team an interface for testing and viewing that activity.
Additional capabilities
For application-defined, multi-step execution, the workflow guide covers explicit step order and success or failure transitions. Use it when your feature needs that orchestration.