Aura – a self-hosted AI agent with temporal graph memory, written in Go (github.com)

Aura, a newly launched self-hosted AI agent, is making waves in the AI/ML community with its unique features and capabilities. Developed in Go, Aura focuses on providing a local-first, provider-neutral platform designed for multi-user environments. It allows users to manage a personal AI agent with tools for document retrieval, scheduled tasks, and integrated memory through a web cockpit, all while ensuring that their data stays local and private. This flexibility enables users to select their AI models—like ChatGPT or local LLMs—without compromising on performance or security. The significance of Aura lies in its innovative use of temporal graph memory, allowing the agent to remember facts, sources, and validity over time, thus enhancing its effectiveness in ongoing tasks. With capabilities like a built-in scheduler, document indexing, and multi-channel support (including CLI and Telegram), Aura not only streamlines user interactions but also provides a robust framework for adaptive reasoning and tool integration. Testing and continuous integration processes are meticulously implemented, ensuring high reliability and performance, making Aura a compelling option for users seeking an efficient and private AI assistant.