# Migrate from Hermes to an open-source AI Operating System

> Hermes is a strong personal agent, and most of what you built there still matters. Moving a team to Kortix, the open-source AI Operating System, is about where the configuration and memory live and who approves a change.

URL: https://hermesalternative.com/migrate-from-hermes

## What the move lands in

`kortix.yaml`

```yaml
# kortix.yaml
agents:
  support-triage:
    file: agents/support-triage.md   # what it does, in plain markdown
    connectors: all                  # what it may reach
```

Source: https://github.com/kortix-ai/suna

An agent and its governance in one reviewable file. Session work reaches `main` through a change request.

## What actually moves

A move from a personal Hermes setup to a company platform is not a rewrite. Three things change hands: the prompts and skills, the memory and context, and the connectors. The governance layer goes in last, once the work already runs.

Kortix is the open-source AI Operating System: your agents, their skills, your company memory, and every connector in one git repo you own, with the agents working on real cloud computers.

## Skills and prompts

Hermes builds skills from experience and stores them as files; [Hermes on GitHub](https://github.com/NousResearch/hermes-agent) documents the learning loop, and it is compatible with the agentskills.io open standard. In Kortix an agent is `agents/<name>.md` (behavior) plus an `agents.<name>` block in `kortix.yaml` (governance). The behavior file is markdown, so a prompt's body moves with edits. The governance block is the new part: which connectors the agent may reach and which secrets it may use.

## Memory and context

Hermes keeps a deepening model of the person it serves, curated across sessions with periodic nudges. Kortix company memory is the living company brain, plain files today, held in the same repo as the agents. That is the shift from one operator's context to shared context: memory becomes something the team reviews and versions rather than something one agent accumulates alone.

## Connectors

Hermes ships 40+ tools behind its own toolset system. Kortix reaches 3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP. Connector credentials are brokered server-side and never enter the machine.

Re-point each connector once. The connector catalog covers the messaging and work apps a team already runs, so a Hermes tool usually has a home in the same list.

## The governance you gain

This is the part Hermes was not built for, and it is the reason a team makes the move. Per-resource permissions for people and agents. Roles, groups, and an audit trail. Approval gates you set. Off until you set them. Session work reaches `main` through a change request, and merge is default-deny for agents.

Identity comes with it: SAML 2.0 single sign-on and SCIM 2.0. Secrets are encrypted at rest with a key per project. One isolated sandbox per session, each session with its own isolated machine and branch. Work scales with the team: thousands of agents in parallel on one config, each on its own cloud computer.

## What stays with Hermes

Nothing about the move requires giving Hermes up. It is a fine personal agent: run it on a $5 VPS or a GPU cluster, talk to it from Telegram while it works on a cloud VM, switch models with `hermes model`. A personal agent and a governed company platform are different jobs. Keep both if the work calls for it.

## Questions about the move

Short answers before a team switches.

### Can we run Hermes and Kortix at the same time?

Yes. Hermes is a personal agent for one operator; Kortix is a platform for a team. Nothing about the move stops anyone using Hermes for their own work.

### Do our agents have to be rewritten?

Most of a prompt carries over. An agent is `agents/<name>.md` (behavior) plus an `agents.<name>` block in `kortix.yaml` (governance); the behavior file is markdown, and the governance block is new.

### Where does our memory end up?

Company memory is plain files in the git repo you own, versioned and diffable, beside the agents.

### What should we move first?

Prompts and skills, then connectors, then the governance layer, once the work already runs.

## Move one team first, then the rest.

Start with one job and one repo you own.

[Try Kortix](https://kortix.com) · [Read the docs](https://kortix.com/docs)

Open source · Any model, your keys · Self-host, VPC, or on-prem