All plugins

netmind-provider

NetMind OpenAI-compatible model provider with a bundled model catalog. Disclosure — registers one model provider at import and nothing else (no tools, hooks, commands or threads); chat requests go to https://api.netmind.ai/inference-api/v1 through Hermes' own client with NETMIND_API_KEY, which Hermes reads; the plugin itself opens no connections, reads and writes no files and starts no processes.

NetMind-AIModelsv1.0.0Updated 26 days ago

About

NetMind OpenAI-compatible model provider with a bundled model catalog. Disclosure — registers one model provider at import and nothing else (no tools, hooks, commands or threads); chat requests go to https://api.netmind.ai/inference-api/v1 through Hermes' own client with NETMIND_API_KEY, which Hermes reads; the plugin itself opens no connections, reads and writes no files and starts no processes.

NetMind Provider for Hermes Agent

This plugin adds NetMind's OpenAI-compatible inference API to Hermes Agent.

Install

This independent plugin has no additional runtime dependencies beyond Hermes.
It requires a host supporting providers.base.ProviderProfile and discovery of
kind: model-provider plugins. Install this local Git checkout with:

hermes plugins install file:///absolute/path/netmind-hermes-provider

After publishing the repository, use its URL and full commit SHA:

hermes plugins install https://github.com/NetMind-AI/netmind-hermes-provider --ref FULL_COMMIT_SHA

Replace FULL_COMMIT_SHA with the full reviewed commit hash from the repository.
Direct repository installation and admission to Hermes' curated catalog are
separate steps. Review the installer's scan and consent prompts.

Configure

Set the NetMind API key in the active Hermes profile's environment:

export NETMIND_API_KEY="your-api-key"

Select the provider and bundled model:

hermes model --provider netmind

You can also run a one-shot request:

hermes -z "Reply with exactly: ok" \
  --provider netmind \
  -m Qwen/Qwen3.8-Flash

Built-in Models

The plugin bundles the 31-model NetMind catalog and uses
Qwen/Qwen3.8-Flash as its default auxiliary model.

The plugin deliberately uses a static catalog and disables model-list discovery
and the corresponding health check. Authenticated /models support has not
been verified; this does not imply NetMind has no such endpoint.

Test

Run the contract tests with a Hermes Agent checkout:

HERMES_AGENT_ROOT=/path/to/hermes-agent \
PYTHONPATH=/path/to/hermes-agent \
python -m pytest -q tests/

The live smoke test also requires NETMIND_API_KEY and NETMIND_LIVE_TEST=1.

Live tests make billable requests. Inference sends prompts, supplied context,
and tool definitions to api.netmind.ai. Keys remain in the Hermes profile
environment and must never be committed.

Catalog submission and maintenance

Hermes' curated catalog requires a public plugin repository, a full 40-character
commit pin, accurate capability declarations, plugin validation including the
install security scan, and a maintainer-reviewed PR. Each subsequent catalog
pin update is a new reviewed PR. This plugin does not update its own code.

Before submission, use hermes plugins validate --help to run the installed
host's validation command. See the authoritative catalog requirements:
https://github.com/NousResearch/hermes-agent/tree/main/plugin-catalog

The source of model metadata is netmind-agent-provider-integrations. In that
repository, run make generate and
python3 scripts/sync_upstreams.py --runtime hermes, then run this plugin's
tests, update its version and publish the reviewed commit.

README from the reviewed commit 00ab432; it updates when the author re-pins.