
Aura AI Agent Framework
by mezmo
Production-ready framework for composing AI agents via TOML with built-in MCP tool integration and RAG pipelines.
What it does
Aura is a high-performance Rust-based framework that allows developers to define AI agents declaratively using TOML configurations. It acts as an orchestrator that integrates LLMs with MCP tools across various transports (HTTP, SSE, STDIO) and connects them to vector stores for RAG capabilities. Aura exposes an OpenAI-compatible web API, making it easy to plug into existing AI clients.
Tools
As a framework, Aura provides the infrastructure to discover and execute tools from other MCP servers. It includes:
- Dynamic MCP Tool Discovery: Automatically discovers tools via configured transports.
- Schema Sanitization: Transforms MCP schemas to ensure compatibility with OpenAI function-calling.
- Multi-Provider Routing: Supports OpenAI, Anthropic, Bedrock, Gemini, and Ollama.
Installation
Clone the repository and build using Cargo:
# Install Rust first, then:
cargo build --release
Configure an agent in config.toml and run the server:
cargo run --bin aura-web-server
Supported hosts
claudechatgpt
Quick install
cargo build --releaseInformation
- Pricing
- free
- Published
- 4/28/2026
- stars






