Hermes: a practical decision dossier
A self-managed assistant with a CLI, messaging gateway and inspectable memory files.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →A personal agent from Nous Research, focused on ongoing use and reusable skills.
A self-managed assistant with a CLI, messaging gateway and inspectable memory files.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →Inventory the real runtime, save a recoverable state and test the seam after restart. Package installation is only one checkpoint.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →Collect and deduplicate deterministically, then use an agent only where interpretation improves the output. Keep the original link and date throughout.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →A local model is one provider option. Separate model access from the assistant runtime, and keep a non-model fallback for collection tasks.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →Validate a changed next action in a fresh session. Storage, retrieval and effective behavior are separate checkpoints.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →Count subscriptions, inference, computers, channels and failed retries separately. Included usage is an allowance, not a guarantee of zero marginal cost.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →Define approvals around concrete actions, recipients and limits. A conversational promise alone does not establish an enforcement boundary.
Check the original: editorial review is due or pending.
Read facts, interpretation and next checks →No English summary yet — see the cited source.
No English summary yet — see the cited source.
No English summary yet — see the cited source.
The official release notes state that Hermes Agent v0.18.1 is a patch release that rolls up approximately 660 PRs merged since v0.18.0, including bug fixes, hardening, and feature work. This version provides a stable release for downstream consumers, such as Docker images, hosted deployments, and PyPI installs. The update can be done via `hermes update` or `pip install -U hermes-agent`
The official release of Hermes Agent v0.18.0, where the team has focused on resolving all P0 and P1 issues and PRs, closing 949 issues and with contributions from 370+ community contributors. This version also introduces new features such as Mixture-of-Agents, model reasoning visualization, work verification, completion contracts, and learning capabilities.
The official release of Hermes Agent v0.17.0 has been announced, featuring new support for iMessage and the Raft agent network, as well as significant enhancements to the desktop application. Additionally, subagents can now run in the background, the image generation tool can edit existing images, and Cursor's Composer model can be accessed through an xAI Grok subscription. This version also includes security updates and fixes for over 300 issues.
The official release of Hermes Agent v0.16.0 is announced, featuring a brand-new native desktop application for macOS, Linux, and Windows, with self-update, drag-and-drop file, and inline model picker capabilities. The web dashboard has also been enhanced with a full browser-based administration panel and quick setup functionality. This version fixes 2 P0 and 62 P1 issues, as well as some security problems.
The official release of Hermes Agent v0.15.2 (v2026.5.29.2) is announced, with bug fixes including packaging issues. The release date is May 29, 2026, and thanks are given to all contributors.
After months with OpenClaw, Joseph Fluckiger compares Hermes for personal operations and visible tooling with OpenClaw's specialist roles and channel organisation. He retains both for different tasks. This is a dated, configuration-specific assessment rather than a universal performance ranking.