Report Abuse

Basic Information

Aurite Agents is a Python framework for building, testing, and running AI agents. It is configuration driven and designed to orchestrate sophisticated agent workflows that interact with external tools, prompts, and resources through the Model Context Protocol (MCP). The repository supports a hierarchical Workspace -> Project organization managed by a .aurite file to share configurations and separate concerns. Users declare components such as agents, LLMs, MCP servers, linear workflows, and custom Python workflows in JSON or YAML. The framework provides an interactive CLI, two textual user interfaces for chat and configuration editing, and a FastAPI REST server for programmatic access and UI development. A frontend ecosystem with a TypeScript API client and a React based Aurite Studio enables visual management, real time streaming, and integration into web applications. Documentation, examples, and templates are included to help users install the package or run the framework from source.

Links

Categorization

App Details

Features
Hierarchical configuration with a Workspace and Project system for shared settings and separation of concerns. Declarative components allow agents, LLMs, MCP servers, tools, linear workflows, and custom Python workflows to be defined in JSON or YAML. An interactive CLI named aurite and two TUIs support initialization, interactive chat with agents, and live configuration editing. Extensible tooling is provided via the Model Context Protocol to connect to local or remote MCP compatible servers and arbitrary tools. Flexible orchestration supports linear_workflows for sequential tasks and custom_workflows for complex Python logic. A FastAPI based REST API exposes framework functionality for programmatic control. Frontend packages include a TypeScript API client with type safety and streaming support and a React Aurite Studio for visual design and monitoring. Integrations include multiple LLM providers via LiteLLM and deployment options such as Docker and cloud platforms.
Use Cases
Aurite speeds creation and operation of agentic applications by providing reusable, declarative building blocks and a clear project hierarchy that simplifies sharing and maintenance. It lets developers configure agents, attach LLMs, and integrate external tools through MCP without rewriting integration code. Workflows can be composed declaratively or implemented in Python for greater control, enabling both simple chains of agents and complex orchestrations. The CLI, TUIs, and REST API support development, testing, and execution across terminal and web environments. The TypeScript API client and Aurite Studio enable integration into web frontends and visual monitoring of live executions and streaming responses. Included documentation, examples, and community registries lower the barrier to adopting LLM providers and MCP tools, and supported deployment options aid moving projects from prototypes to production.

Please fill the required fields*