Hermes Mesh
B1Z0N/hermes-mesh
Keeps skills and durable memory in sync across every machine that runs a Hermes agent
Hermes Mesh is a set of shell and Python scripts that syncs Hermes skills, agent memory and user profile between machines through a shared Git repository. It runs every 15 minutes in the background.
What Hermes Mesh does
Hermes Mesh addresses a specific problem: Hermes agents on a VPS, a laptop and a work PC each build different memories. Mesh keeps skills, durable agent memory and the user profile in sync, so the agent remembers the same things on every machine. One machine acts as the coordinator and holds a bare Git repository, and the others are workers that need SSH key access to it.
Each machine pulls, merges its own additions and pushes back every 15 minutes, using a cron job on Linux or launchd on macOS. Three-way merges handle conflicts, and the README says an LLM resolves two machines editing the same fact differently. Facts can be tagged with a machine name so hardware specs and local paths stay on the right machine. Files such as .env, auth.json, state.db, sessions, logs and cron stay local. The repository includes setup, sync, update, test and uninstall scripts and a --dry-run option.
Key features
- Syncs skills, agent memory, user profile and the scripts themselves
- Machine tags keep hardware facts and local paths on one machine
- Three-way merge, with LLM help for conflicting edits
- Scheduled sync every 15 minutes through cron or launchd
- Secrets, sessions, logs and
.envfiles never leave their machine - sync.sh options for --dry-run, --force-push, --force-pull and --resolve-skills
When to use it
- Share one agent memory between a VPS and a MacBook
- Make skills created on one machine available on every other Hermes install
- Preview a sync with --dry-run before changing anything
Who it is for: Hermes Agent users who run the agent on more than one machine and want one consistent memory.
How it fits with Hermes Agent
It is built specifically for Hermes Agent and works on the ~/.hermes skills and memories directories.
How to install Hermes Mesh
These commands are copied from the project's README. Check the repository for the latest steps before you run them.
curl -sSL https://raw.githubusercontent.com/B1Z0N/hermes-mesh/main/setup.sh | bashRequirements: Python 3.8+ (with tomli below 3.11), Git, SSH key access from workers to the coordinator, and cron on Linux or launchd on macOS
FAQ
What is Hermes Mesh?
Hermes Mesh is a tool that syncs Hermes Agent skills and durable memory across machines through a shared Git repository. It runs on a schedule every 15 minutes.
How do I install Hermes Mesh?
Run the one-line setup script from the README on each machine and answer its questions. The README also suggests pinning the script URL to a specific commit SHA if you want extra assurance.
What stays local and does not sync?
Files such as .env, auth.json, state.db, sessions, logs and cron stay on their own machine, as do memory entries tagged for one machine. Secrets, OAuth tokens and API keys never sync.
Similar memory for Hermes Agent
All memoryTyped long-term memory for AI agents, as an MCP server and a Hermes plugin
kigner Multimodal WikiHermes skill that builds an interlinked markdown wiki from text, web pages, PDFs, screenshots and audio
skynetcmd m3 MemoryLocal-first shared memory for AI agents with hybrid search, MCP tools and a fully offline mode
TechieTer Hermes Memory Keep-Alive for ObsidianObsidian task notes and a stall-recovery loop for Hermes agents
Ladybug-Memory Hermes Ladybug MemoryLocal memory provider for Hermes Agent backed by an embedded Ladybug graph database
electricsheephq LCM-XLossless context memory for Hermes Agent using a SQLite-backed summary DAG
Related guides: SOUL.md for Hermes Agent: what it is and how to write one · Run multiple Hermes agents with profiles