# Honcho-AIO

Honcho-AIO is a self-hosted ai app. Honcho is self-hosted memory infrastructure for AI agents that stores messages, derives context and peer conclusions in the background, and serves context queries, search, and historical recall over people, sessions,... It is open source and self-hosted, installable on Docker Hub and Unraid.

Canonical page: https://ourlocalhost.com/app/2419/honcho-aio

## Description

Honcho is self-hosted memory infrastructure for AI agents that stores messages, derives context and peer conclusions in the background, and serves context queries, search, and historical recall over people, sessions, and ideas. honcho-aio packages the full stack—Honcho API, PostgreSQL 17 with pgvector, Redis, and s6 supervisor—into a single Unraid container, keeping all databases on loopback so nothing is exposed to your LAN. It provides a hosted memory backend for agent harnesses such as Hermes, letting you give agents persistent memory without running separate database containers.

## Platform availability

Available on 1 of 10 platforms: Unraid.

## Health score

49/100 (Fair), based on 3 of 4 factors.
- Maintained: 86 (last commit 15 days ago)
- Popular: 0 (0 GitHub stars)
- Easy to install: 35 (1 install platform, install notes, Docker image)

## Links

- Docker Hub: https://hub.docker.com/r/dub19/honcho-aio
- GitHub: https://github.com/wgross19/honcho-aio
- Unraid: https://ca.unraid.net/apps/honcho-aio-1jt5uz71ri6hzz?q=honcho-aio

## Common questions

### Why does honcho-aio exit with code 64 on first boot?

It refuses to start if POSTGRES_PASSWORD is empty. Set a PostgreSQL password before the first boot.

### Are the Postgres and Redis services reachable from the LAN?

No. PostgreSQL and Redis are loopback-only and are never published. Agents connect only through the Honcho HTTP API on port 8000.

### How do I configure the embedding and chat models for local or cloud use?

Set EMBEDDING_MODEL and CHAT_MODEL as provider/model strings, splitting on the first slash. For local/, set LOCAL_BASE_URL to the OpenAI-compatible endpoint, and local chat has no cloud fallback. Boot fails if the embedding width does not match EMBEDDING_VECTOR_DIMENSIONS; the default is 768 for embeddinggemma.

### How are Honcho versions managed in honcho-aio?

The container is pinned to official plastic-labs/honcho tags. It uses HONCHO_VERSION and HONCHO_GIT_SHA to select the pinned build.

---
Source: Our Local Host (https://ourlocalhost.com/app/2419/honcho-aio), the self-hosted and open-source app directory.
