Ollama has announced the launch of Hermes Agent, a self-improving AI agent developed by NousResearch, as part of Ollama version 0.21. This new feature allows users to deploy Hermes Agent completely offline, ensuring data privacy while retaining full functionality without relying on cloud services. Additionally, Hermes Agent addresses the challenge of ‘AI amnesia’ by transforming learned information into reusable skills that improve with experience, enhancing the user experience across various messaging platforms such as Telegram, Discord, Slack, WhatsApp, and Signal.

Ollama: Ollama is a platform that enables users to run and manage large language models locally on their machines. The platform integrates with various AI agents and frameworks, as demonstrated by its support for Hermes Agent in version 0.21, allowing seamless deployment of self-improving AI systems with customizable model selection and multi-platform connectivity.
Nous Research: Nous Research is an AI research organization that develops open-source AI agents and models. The company created Hermes Agent, a self-improving AI agent framework that addresses persistent memory challenges in conversational AI and includes automatic skill creation, cross-session memory retention, and multi-platform integration capabilities.

`json
{
“Local deployment”: “Users can run Hermes Agent entirely offline using Ollama with local models, ensuring data privacy while maintaining full agent functionality without cloud dependencies.”,
“Self-improving capability”: “Hermes Agent transforms learned information into reusable skills and improves them through experience, addressing the issue of AI systems losing memory between interactions.”
}
`