42-evey

evey-cache

Delegation caching plugin for Hermes Agent that avoids re-asking the same questions to models.

290 stars31 forksPythonMIT

Repository last updated Jul 18, 2026

Overview

The Evey Cache plugin provides delegation caching for Hermes Agent, preventing redundant model calls by caching responses to repeated queries. This reduces latency and cost when agents ask the same questions multiple times.

When an agent delegates a task to a model, the plugin checks if a similar request has been made before. If a cached response exists, it returns that instead of invoking the model again. This is especially useful in multi-turn conversations or when multiple agents share the same knowledge base.

The plugin is designed for developers using Hermes Agent who want to optimize performance and minimize API costs. It integrates seamlessly as a custom plugin, requiring minimal configuration.

Highlights

  • Caches model responses to avoid redundant queries
  • Reduces latency and API costs for repeated questions
  • Seamless integration with Hermes Agent delegation
  • Simple plugin setup with minimal configuration

Features

Delegation Caching

Automatically caches responses from model delegations and serves cached results for identical queries.

Cost Optimization

Reduces the number of API calls to models, lowering operational costs for agent workflows.

Performance Boost

Speeds up response times by avoiding repeated model invocations for common questions.

Installation

Commands are reproduced exactly as published by the project maintainers. Always check the upstream repository for the current instructions.

git clone https://github.com/42-evey/hermes-plugins.git
cp -r hermes-plugins/evey-* ~/.hermes/plugins/
cp hermes-plugins/evey_utils.py ~/.hermes/plugins/
# Restart hermes-agent

Requirements

  • Hermes Agent installed
  • Python 3.x

How it's built

The plugin is implemented in Python and follows the Hermes Agent plugin structure. The repository contains a single plugin directory evey-cache/ with an __init__.py file and a plugin.yaml manifest. The plugin.yaml defines the plugin metadata and entry points.

Integration requires placing the plugin in the Hermes Agent plugins directory and ensuring the agent is configured to use delegation caching. The plugin intercepts delegation requests and checks a cache store (likely in-memory or configurable) before forwarding to the model.

Source

This summary was written for Neura Market from the project's own repository. Install commands, identifiers and licence details are reproduced unchanged. The 42-evey/evey-cache repository is the authoritative source and the place to file issues or contribute.

Category: Cost & Governance · Licensed under MIT

Related Hermes plugins