Profile
- Memory stack that actually compounds — Start with built-in Hermes memory, then add honcho-self-hosted when you want stronger cross-session user modeling, hindsight when you need retain/recall/reflect workflows across large histories, and plur when you want portable shared memory artifacts in an open engram format. If you also want proactive recall, pair it with flowstate-qmd.
Signals
Listed in the awesome-hermes-agent README
Sources: 2 / Surfaces: 1
What the upstream surface says
Short excerpt only, so you can decide whether to click out.
Self-host Honcho (Plastic Labs' memory layer) on your own server instead of using their cloud. Works with Hermes Agent out of the box.
No fork required — just 3 config files on top of upstream Honcho.
Hermes Agent has a 4-layer memory system. The cross-session memory layer is powered by Honcho, which builds a deepening model of the user across conversations — extracting observations, recalling context, and consolidating memories over time.
- Runs Honcho's full memory stack (API, Deriver, PostgreSQL, Redis) on your machine
- Routes LLM calls through any OpenAI-compatible provider (primary + backup)
- All your data stays on your machine — no third-party cloud storage
- Works with OpenRouter, Venice, Routstr, Together, Ollama, or any other provider
- Ubuntu 22.04+ (VM, VPS, bare metal, or any Linux server — tested on 22.04, 6GB RAM, 80GB disk)
- Docker Engine + Compose plugin
- API key from any OpenAI-compatible provider (openrouter.ai, venice.ai, together.ai, etc.)
- Second API key for backup (optional)