A cloud runtime for Hermes that preserves the config-driven workflow
Hermes is a general-purpose autonomous agent with persistent memory, self-improving skills, and multi-platform messaging. It runs on Python and uv, configured through YAML and dotfiles. That setup works well on a single server. The runtime question is how to move that same agent into a managed, isolated cloud environment without rewriting the configuration.
What Makes Hermes Different
A Python-native agent with its own runtime needs
Hermes is not a Node harness. Its stack, config model, and persistent memory system shape what the runtime needs to provide.
Native Python runtime support
Hermes runs on Python, not Node. The runtime must support Python toolchains and dependency management natively, not treat them as an afterthought.
Config and secrets managed for you
Hermes keeps its config and secrets in local files. The runtime projects these into the environment automatically. No manual file placement or setup scripts per run.
CLI or API, same environment either way
Run single prompts for scripted tasks, or start the API server for programmatic access via OpenAI-compatible endpoints. Both modes share the same persistent environment underneath.
Persistent memory needs a persistent environment
Hermes remembers past interactions and refines its skills over time. That persistent memory only works if the environment persists too. Ephemeral sandboxes erase everything the agent has learned.
From Local to Cloud
How Hermes moves from a single server to a managed environment
Hermes is designed to run on your own infrastructure. Install it, configure your model, and run it from the terminal. That self-hosted model is fast and self-contained.
The gap emerges when you need the same agent to run in a team context, in CI, or on a schedule. Config files need to exist somewhere persistent. Secrets need proper projection instead of sitting in plaintext dotfiles. The Python environment and installed packages need to survive between runs so you are not reinstalling dependencies every time.
On Spinup, Hermes runs inside an isolated agent environment. The control plane loads configuration automatically and projects secrets rather than exposing them in dotfiles. The Python environment and installed packages persist across runs. The result is the same Hermes workflow, running in a managed environment with the controls a team actually needs.
This also means the Hermes harness stays portable. If you later want to compare Hermes with Claude Code or OpenClaw on the same task, the agent identity, environment, and secrets remain stable. You swap the harness, not the entire runtime.
FAQ
Hermes runtime questions
How does Hermes configuration work in a cloud runtime?+
Hermes keeps its config and secrets in local files. On Spinup, this configuration maps cleanly into the agent environment. The control plane loads config automatically and projects secrets instead of storing them in plaintext dotfiles.
Can I use both Hermes CLI mode and the API server on Spinup?+
Yes. Run single prompts for scripted tasks, or start the API server for programmatic access via OpenAI-compatible endpoints. Both modes work inside a Spinup environment. The runtime handles Python process lifecycle and filesystem persistence in either mode.
How does Hermes being open-source affect runtime choice?+
Hermes is built by Nous Research and developed in the open. Its Python/uv stack, config model, and execution modes are community-driven. On Spinup, Hermes plugs in as one harness among several. If community direction shifts or you want to compare harnesses, your agent environment, secrets, and snapshots stay stable above the change.
Why not just self-host Hermes?+
Hermes works well self-hosted on a VPS or server: config files in the home directory, uv for dependency management, and CLI-driven workflows. The gap appears when you need managed isolation, team-level controls, or the ability to compare harnesses. Spinup bridges that gap by giving the same agent a managed cloud environment with persistent state and controlled secrets, without changing the Hermes workflow itself.
Hermes in the Spinup Runtime
Hermes is one harness inside a broader runtime model
The agent runtime defines the environment, controls, and lifecycle above the harness layer. Hermes plugs into that model alongside other harnesses.
Early Access
Keep the Hermes workflow. Add the runtime controls around it.
Join the early-access waitlist if this is the runtime shape your team has been missing.