A self-improving AI agent application that creates and refines skills from experience while maintaining context across sessions. It runs from a $5 VPS to a GPU cluster and connects to multiple messaging platforms.
Project overview
A self-improving AI agent application that creates and refines skills from experience while maintaining context across sessions. It runs from a $5 VPS to a GPU cluster and connects to multiple messaging platforms.
Project type
AI Agent
Use cases
Automation
Deployment
Refer to project documentation
License
MIT
Best for
Developers and general users seeking an AI agent that maintains a persistent user model and improves its own skills autonomously.
Operators who need cross-platform messaging continuity alongside scheduled automation on low-cost or serverless hardware.
Key capabilities
Agent-curated memory with periodic nudges, autonomous skill creation after complex tasks, and skills that self-improve during use.
Full TUI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output.
Single gateway process delivering the agent to Telegram, Discord, Slack, WhatsApp, Signal, or Email with cross-platform conversation continuity.
Built-in cron scheduler with delivery to any platform for unattended daily reports, nightly backups, or weekly audits.
Spawn isolated subagents for parallel workstreams and write Python scripts that call tools via RPC to collapse multi-step pipelines.
Six terminal backends — local, Docker, SSH, Singularity, Modal, and Daytona — including serverless persistence that hibernates when idle.
Limitations and risks
Antivirus software such as Bitdefender or Windows Defender may flag the bundled uv.exe as malware due to a false positive.
Telemetry practices and data boundary policies are currently not documented in the available facts.
Getting started
Setup is rated as easy because automated install scripts handle dependencies like uv, Python 3.11, and Node.js. Coding knowledge is not required.
Users can get started by running the install script, followed by executing the 'hermes setup' command via the command-line interface.
Evidence and sources
README: The self-improving AI agent built by Nous Research. It's the only agent with a built-in learning loop — it creates skills from experience, improves them during use, nudges itself…
README: The installer handles everything: uv, Python 3.11, Node.js, ripgrep, ffmpeg, and a portable Git Bash
README: If your antivirus (Bitdefender, Windows Defender, etc.) quarantines `uv.exe` from the Hermes `bin` folder (`%LOCALAPPDATA%\hermes\bin\uv.exe`), this is a false positive.
README: computer-use-linux — Linux desktop-control MCP server for Hermes and other MCP hosts