Persistent Hermes on Hugging Face Spaces

HermesFace packages Hermes as a Docker-based Hugging Face Space with a browser dashboard and messaging gateway, periodically syncing /opt/data to a private Dataset repository so conversations, skills, memory, and configuration survive container restarts.

Hermes Intermediate Copied: 0
1 resource

Requirements

  • The image passes supported Hermes environment variables through to the agent, so model providers and messaging channels can be configured with Space secrets.
  • A Hugging Face account and Space with a write-capable token.
  • A HF_TOKEN secret and one model-provider key must be added to the Space.
  • Hugging Face account, Space, and write-capable token.
  • A private Dataset repository for persistence.
  • At least one supported model-provider credential.
  • Optionally, a Telegram, Discord, Slack, WhatsApp, or another Hermes gateway credential.

How to use

Hugging Face quotas, sleep behavior, hardware allocations, and pricing can change, so treat free-tier and always-on claims as platform-dependent rather than permanent guarantees.

Keep the persistence Dataset private, as it may contain full conversations, memory, custom skills, and configuration.

Secrets remain server-side, but the agent still has the network and tool capabilities allowed by the Space container and Hermes configuration.

  1. Duplicate the public HermesFace Space.
  2. Create a Hugging Face token with write permission.
  3. Add HF_TOKEN and one model-provider key under Settings → Repository secrets.
  4. Set AUTO_CREATE_DATASET=true, or create a private Dataset and set HERMES_DATASET_REPO explicitly.
  5. Update the duplicated Space's datasets: metadata so it does not point at the original author's dataset.
  6. Open the Space URL and configure messaging through the dashboard or environment variables. For manual persistence, the documented settings include:
AUTO_CREATE_DATASET=true
SYNC_INTERVAL=60
TZ=UTC

HermesFace changes deployment and persistence, not Hermes's prompt model. Use the same tasks you would use locally, then restart the Space and verify that the relevant session, memory, or skill returns from the private Dataset backup.

Source: aliaihub/awesome-hermes-usecases and its contributors. License: CC-BY-4.0. Source license notice.

Translation and editorial adaptation by ClawCases.

>Resources

>Related Use Cases

>More at this level: Intermediate