apify-hermes-agent-plugin

Hermes Agent plugin for Apify, the world's largest marketplace for AI tools - discover Actors in Apify Store, start runs, and collect structured results, all from natural language, via Hermes Agent.

Install

pip install apify-hermes-agent-plugin
hermes plugins enable apify

Then run:

hermes apify-setup

This prompts for your APIFY_API_TOKEN (get one at https://console.apify.com/settings/integrations?utm_source=hermes-agent&utm_medium=integrations), saves it, and enables the Apify Actors toolset for the CLI. Pass --token <token> to skip the prompt.

Troubleshooting: hermes plugins enable apify fails

If that command prints Plugin 'apify' is not installed or bundled., your installed hermes-agent predates 0.18.1, when entry-point plugin discovery was fixed upstream. Check your version and upgrade first:

hermes --version
pip install --upgrade hermes-agent

If you can't upgrade, enable the plugin manually instead: open ~/.hermes/config.yaml and add apify to the plugins.enabled list:

plugins:
  enabled:
    - apify

Then run hermes apify-setup as above - it still takes care of enabling the toolset for you.

Tools

• apify_discover - search Apify Store by keyword, or fetch an Actor's input schema + README by actor_id.
• apify_start - fire-and-forget batch Actor starts (up to 10 per call).
• apify_collect - poll run statuses and return completed dataset results.

Web search & fetch

This plugin also registers apify as a Hermes Agent web search and web-fetch backend, running the RAG Web Browser Actor for search and the Web Fetch Actor for content extraction. Both use the same APIFY_API_TOKEN configured above - no separate setup needed.

To use them, set the search and/or extract backend in ~/.hermes/config.yaml:

web:
  search_backend: apify
  extract_backend: apify

or select apify interactively via hermes tools.

Development

pip install -r requirements-dev.txt
pytest

Releasing

Releases are cut from the Actions tab, not by creating a GitHub Release by hand: open the publish workflow → Run workflow → choose a release type (auto, patch, minor, major, or custom). The pipeline then bumps the version, updates CHANGELOG.md, creates the GitHub Release, and publishes to PyPI automatically.

auto infers the version bump from Conventional Commits since the last tag, and only works correctly when at least one qualifies (feat:, fix:, etc. — see AGENTS.md's Release process for the exact rules). If you're not sure, pick patch/minor/major explicitly instead — those always compute correctly regardless of commit history.