A self-improving AI agent that creates and refines skills from experience, remembers context across sessions, and builds a deepening user model. It can be run from a $5 VPS to a GPU cluster and is accessible via terminal or multiple messaging platforms.
Project overview
The agent features a learning loop that autonomously creates and improves skills based on past tasks and user interactions, maintaining context across sessions.
Project type
AI Agent · AI Coding
Use cases
Coding & Development
Deployment
Refer to project documentation
License
MIT
Best for
Developers and general users seeking a persistent agent that autonomously creates skills, remembers context, and can be delivered across messaging platforms.
Key capabilities
Provides agent-curated memory with periodic nudges, autonomous skill creation after complex tasks, and skills that self-improve during use.
Features a full TUI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output.
A single gateway process delivers the agent to Telegram, Discord, Slack, WhatsApp, Signal, or Email with cross-platform conversation continuity.
Built-in cron scheduler delivers to any platform for unattended daily reports, nightly backups, or weekly audits.
Spawns isolated subagents for parallel workstreams and writes Python scripts that call tools via RPC to collapse multi-step pipelines.
Supports six terminal backends, including local, Docker, SSH, Singularity, Modal, and Daytona, with serverless persistence that hibernates when idle.
Limitations and risks
Antivirus software like Bitdefender or Windows Defender may flag the bundled uv.exe as malware during installation.
The data boundary and telemetry practices are not documented.
Getting started
Setup difficulty is rated as easy because automated install scripts handle uv, Python 3.11, Node.js, and other dependencies. The first success path is to run the install script, followed by running hermes setup.
Alternatives and comparisons
Provides a server component where deployed agents run continuously and can be triggered by external sources.
Build and operationalize AI workflows and multi-step agents using your own data, models, and tools.