feat: add Honcho memory provider with PostgreSQL + pgvector stack

Add Honcho (https://github.com/plastic-labs/honcho) as a self-hosted memory infrastructure for stateful AI agents. Changes: - ai/honcho/Dockerfile: multi-stage build from Honcho GitHub source - ai/honcho/init.sql: CREATE EXTENSION vector for pgvector - ai/compose.yml: add honcho-db (pgvector/pgvector:pg17-trixie) and honcho services with ai_backend/ai_net networking and Traefik labels - build/honcho/config.toml: pre-configured for Ollama embeddings (nomic-embed-text via http://ollama:11434/v1), deriver/summary/dream disabled by default - env/.env.example.honcho: sample env vars (HONCHO_DB_PASSWORD, LLM_OPENAI_API_KEY) Usage: cp env/.env.example.honcho .env # edit secrets mkdir -p /mnt/HoardingCow_docker_data/Honcho cp build/honcho/config.toml /mnt/HoardingCow_docker_data/Honcho/config.toml docker compose -f ai/compose.yml up honcho
fix: add WORKDIR and httpx dependency to Hermes Dockerfile
2026-05-20 14:19:58 -04:00 · 2026-05-20 14:18:24 -04:00 · 2026-05-20 14:05:45 -04:00
13 changed files with 401 additions and 570 deletions
--- a/ai/compose.yml
+++ b/ai/compose.yml
@@ -32,24 +32,19 @@ services:
        - default
    container_name: hermes
    entrypoint: ["/bin/bash", "-c",
-      "bash /opt/data/hermes-tools/install.sh && bash /usr/local/bin/run-multi-gateways.sh && exec /usr/bin/tini -g -- /opt/hermes/docker/entrypoint.sh \"$@\"",
+      "bash /opt/data/hermes-tools/install.sh && exec /usr/bin/tini -g -- /opt/hermes/docker/entrypoint.sh \"$@\"",
      "hermes-entrypoint"]
    restart: always
    # Gateway run enables the internal API server on port 8642
    command: gateway run
    environment:
      - OLLAMA_HOST=http://ollama:11434
-      - HERMES_DASHBOARD=1
-      # Multi-profile: comma-separated list of profiles to run as gateways.
-      # The entrypoint reads this and starts one gateway per profile.
-      # Add profiles here when they exist on disk (e.g. default,researcher,writer)
-      - HERMES_PROFILES=ashley,claire,finn,matt,paul
      - API_SERVER_ENABLED=true
      - API_SERVER_PORT=8642
      - API_SERVER_HOST=0.0.0.0
      - API_SERVER_KEY=hermes_local_key
      - GATEWAY_ALLOW_ALL_USERS=true
-      - OPENROUTER_API_KEY=${OPENROUTER_API_KEY}
+      - OPENROUTER_API_KEY=${OPEN...KEY}
      # ROCm for GPU-accelerated faster-whisper STT
      - HSA_OVERRIDE_GFX_VERSION=9.0.6
      - HCC_AMDGPU_TARGET=gfx906
@@ -71,32 +66,6 @@ services:
      - "26"
    networks:
      - ai_backend
-      - ai_net
-    depends_on:
-      - honcho
-    labels:
-      - "traefik.enable=true"
-      - "traefik.docker.network=ai_net"
-
-      # Router for HTTP + redirection to HTTPS
-      - "traefik.http.routers.hermes-web-http.rule=Host(`hermes.lazyworkhorse.net`)"
-      - "traefik.http.routers.hermes-web-http.entrypoints=web"
-      - "traefik.http.routers.hermes-web-http.middlewares=redirect-to-https"
-
-      # Router for HTTPS with TLS — protected by Authelia
-      - "traefik.http.routers.hermes-web-https.rule=Host(`hermes.lazyworkhorse.net`)"
-      - "traefik.http.routers.hermes-web-https.entrypoints=websecure"
-      - "traefik.http.routers.hermes-web-https.tls=true"
-      - "traefik.http.routers.hermes-web-https.tls.certresolver=njalla"
-      - "traefik.http.routers.hermes-web-https.middlewares=hermes-auth"
-
-      # Authelia forwardAuth
-      - "traefik.http.middlewares.hermes-auth.forwardauth.address=http://authelia:9091/api/verify?rd=https://auth.lazyworkhorse.net/"
-      - "traefik.http.middlewares.hermes-auth.forwardauth.trustforwardheader=true"
-      - "traefik.http.middlewares.hermes-auth.forwardauth.authresponseheaders=X-Forwarded-User,X-Forwarded-Groups"
-
-      # Service Loadbalancer (dashboard port 9119)
-      - "traefik.http.services.hermes-web.loadbalancer.server.port=9119"

  syncthing:
    image: syncthing/syncthing:latest
@@ -127,7 +96,6 @@ services:
      - "traefik.http.routers.syncthing-https.tls.certresolver=njalla"
      - "traefik.http.services.syncthing.loadbalancer.server.port=8384"

-
  ollama:
    build:
      context: ./ollama
@@ -161,26 +129,109 @@ services:
      - "303"
      - "26"

-  # --- Honcho + OpenConcho combiné: API + Web UI nginx/FastAPI ---
+  paperclip-db:
+    image: postgres:17-alpine
+    container_name: paperclip-db
+    restart: always
+    environment:
+      POSTGRES_USER: paperclip
+      POSTGRES_PASSWORD: ${PAPERCLIP_DB_PASSWORD:?PAPERCLIP_DB_PASSWORD must be set}
+      POSTGRES_DB: paperclip
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U paperclip -d paperclip"]
+      interval: 5s
+      timeout: 5s
+      retries: 10
+    volumes:
+      - /mnt/HoardingCow_docker_data/Paperclip/pgdata:/var/lib/postgresql/data
+    networks:
+      - ai_backend
+
+  paperclip:
+    image: ghcr.io/paperclipai/paperclip:v2026.517.0
+    container_name: paperclip
+    restart: always
+    ports:
+      - "127.0.0.1:3100:3100"
+    environment:
+      - HOST=0.0.0.0
+      - PORT=3100
+      - SERVE_UI=true
+      - DATABASE_URL=postgres://paperclip:***@paperclip-db:5432/paperclip
+      - BETTER_AUTH_SECRET=${PAPE...CRET must be set}
+      - PAPERCLIP_PUBLIC_URL=https://paperclip.lazyworkhorse.net
+      - PAPERCLIP_DEPLOYMENT_MODE=authenticated
+      - PAPERCLIP_DEPLOYMENT_EXPOSURE=private
+    volumes:
+      - /mnt/HoardingCow_docker_data/Paperclip/data:/paperclip
+    depends_on:
+      paperclip-db:
+        condition: service_healthy
+    networks:
+      - ai_net
+      - ai_backend
+    labels:
+      - "traefik.enable=true"
+      - "traefik.docker.network=ai_net"
+
+      - "traefik.http.routers.paperclip-http.rule=Host(`paperclip.lazyworkhorse.net`)"
+      - "traefik.http.routers.paperclip-http.entrypoints=web"
+      - "traefik.http.routers.paperclip-http.middlewares=redirect-to-https"
+
+      - "traefik.http.routers.paperclip-https.rule=Host(`paperclip.lazyworkhorse.net`)"
+      - "traefik.http.routers.paperclip-https.entrypoints=websecure"
+      - "traefik.http.routers.paperclip-https.tls=true"
+      - "traefik.http.routers.paperclip-https.tls.certresolver=njalla"
+
+      - "traefik.http.services.paperclip.loadbalancer.server.port=3100"
+
+  # ---------------------------------------------------------------------------
+  # Honcho — Memory infrastructure for stateful AI agents
+  # Self-hosted memory server with pgvector for embedding storage.
+  # Defaults to Ollama for embeddings; configure LLM provider for full deriver
+  # and summarization support.
+  #
+  # API port: 8000
+  # Web:       https://honcho.lazyworkhorse.net
+  # Docs:      https://github.com/plastic-labs/honcho
+  # ---------------------------------------------------------------------------
+
+  honcho-db:
+    image: pgvector/pgvector:pg17-trixie
+    container_name: honcho-db
+    restart: unless-stopped
+    environment:
+      POSTGRES_DB: honcho
+      POSTGRES_USER: honcho
+      POSTGRES_PASSWORD: ${HONCHO_DB_PASSWORD:?HONCHO_DB_PASSWORD must be set}
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U honcho -d honcho"]
+      interval: 5s
+      timeout: 5s
+      retries: 10
+    volumes:
+      - /mnt/HoardingCow_docker_data/Honcho/pgdata:/var/lib/postgresql/data
+      - ./honcho/init.sql:/docker-entrypoint-initdb.d/init.sql
+    networks:
+      - ai_backend
+
  honcho:
    build:
      context: ./honcho
-      ssh:
-        - default
+      dockerfile: Dockerfile
    container_name: honcho
    restart: unless-stopped
+    ports:
+      - "127.0.0.1:8000:8000"
+    depends_on:
+      honcho-db:
+        condition: service_healthy
    environment:
-      - DB_CONNECTION_URI=postgresql+psycopg://honcho:honcho_pass@honcho-db:5432/honcho
-      - CACHE_URL=redis://honcho-redis:6379/0
-      - CACHE_ENABLED=true
-      - EMBEDDING_VECTOR_DIMENSIONS=1024
-      - AUTH_USE_AUTH=true
-      - AUTH_JWT_SECRET=${HONCHO_AUTH_JWT_SECRET}
-      # Needed by deriver/dream to make LLM calls (api_key_env = "HONCHO_OPENAI_API_KEY" in config.toml)
-      - HONCHO_OPENAI_API_KEY=${HONCHO_OPENAI_API_KEY}
+      DB_CONNECTION_URI: postgresql+psycopg://honcho:${HONCHO_DB_PASSWORD:?HONCHO_DB_PASSWORD must be set}@honcho-db:5432/honcho
+      LOG_LEVEL: INFO
+      LLM_OPENAI_API_KEY: ${LLM_OPENAI_API_KEY:-ollama}
    volumes:
-      - honcho_data:/app/data
-      - /mnt/HoardingCow_docker_data/Honcho/config.toml:/app/config.toml:ro
+      - /mnt/HoardingCow_docker_data/Honcho/config.toml:/app/config.toml
    networks:
      - ai_backend
      - ai_net
@@ -188,64 +239,39 @@ services:
      - "traefik.enable=true"
      - "traefik.docker.network=ai_net"

-      # Router for HTTP + redirect to HTTPS
      - "traefik.http.routers.honcho-http.rule=Host(`honcho.lazyworkhorse.net`)"
      - "traefik.http.routers.honcho-http.entrypoints=web"
      - "traefik.http.routers.honcho-http.middlewares=redirect-to-https"

-      # Router for HTTPS with TLS — protected by Authelia
      - "traefik.http.routers.honcho-https.rule=Host(`honcho.lazyworkhorse.net`)"
      - "traefik.http.routers.honcho-https.entrypoints=websecure"
      - "traefik.http.routers.honcho-https.tls=true"
      - "traefik.http.routers.honcho-https.tls.certresolver=njalla"
-      - "traefik.http.routers.honcho-https.middlewares=hermes-auth"

-      # Service Loadbalancer (nginx port)
-      - "traefik.http.services.honcho.loadbalancer.server.port=80"
-    depends_on:
-      honcho-db:
-        condition: service_healthy
-      honcho-redis:
-        condition: service_healthy
+      - "traefik.http.services.honcho.loadbalancer.server.port=8000"

-  honcho-db:
-    image: pgvector/pgvector:pg15
-    container_name: honcho-db
+  holographic-memory:
+    build:
+      context: ./holographic-memory
+    image: holographic-memory:latest
+    container_name: holographic-memory
    restart: unless-stopped
    ports:
-      - "127.0.0.1:5432:5432"
-    command: ["postgres", "-c", "max_connections=200"]
+      - "127.0.0.1:8100:8100"
    environment:
-      - POSTGRES_DB=honcho
-      - POSTGRES_USER=honcho
-      - POSTGRES_PASSWORD=honcho_pass
-      - PGDATA=/var/lib/postgresql/data/pgdata
+      - HOLOGRAPHIC_DB_PATH=/data/holographic/memory_store.db
+      - HOLOGRAPHIC_PORT=8100
+      - HOLOGRAPHIC_DEFAULT_TRUST=0.5
    volumes:
-      - /mnt/HoardingCow_docker_data/Honcho/postgres:/var/lib/postgresql/data
-      - ./honcho/init-db.sql:/docker-entrypoint-initdb.d/init.sql:ro
+      - /mnt/HoardingCow_docker_data/HolographicMemory:/data/holographic
    networks:
      - ai_backend
    healthcheck:
-      test: ["CMD-SHELL", "pg_isready -U honcho -d honcho"]
-      interval: 5s
+      test: ["CMD", "python3", "-c", "import urllib.request; urllib.request.urlopen('http://127.0.0.1:8100/health')"]
+      interval: 30s
      timeout: 5s
-      retries: 5
-
-  honcho-redis:
-    image: redis:8
-    container_name: honcho-redis
-    restart: unless-stopped
-    ports:
-      - "127.0.0.1:6379:6379"
-    volumes:
-      - /mnt/HoardingCow_docker_data/Honcho/redis:/data
-    networks:
-      - ai_backend
-    healthcheck:
-      test: ["CMD-SHELL", "redis-cli ping"]
-      interval: 5s
-      timeout: 5s
-      retries: 5
+      retries: 3
+      start_period: 10s

 networks:
  ai_net:
@@ -254,11 +280,6 @@ networks:
  ai_backend:
    driver: bridge
    name: ai_backend
-
-volumes:
-  honcho_data:
-    external: true
-    name: honcho_data
    
  # llama_cpp_devstral:
  #   image: ghcr.io/ggml-org/llama.cpp:server-rocm
@@ -403,8 +424,8 @@ volumes:
  #     - /home/gortium/infra:/data/workspace/infra
  #   environment:
  #     - TZ=America/Toronto
-  #     - OPENCLAW_GATEWAY_TOKEN=${OPENCLAW_GATEWAY_TOKEN}
-  #     - OPENROUTER_API_KEY=${OPENROUTER_API_KEY}
+  #     - OPENCLAW_GATEWAY_TOKEN=${OPEN...KEN}
+  #     - OPENROUTER_API_KEY=${OPEN...KEY}
  #     # Point to the sidecar browser
  #     - BROWSER_CDP_URL=http://openclaw-browser:9222
  #     - BROWSER_EVALUATE_ENABLED=true
@@ -449,7 +470,7 @@ volumes:
  #     - PGID=1000
  #     - PUBLIC_KEY_FILE=/config/ssh/authorized_keys
  #     - SUDO_ACCESS=false
-  #     - PASSWORD_ACCESS=false
+  #     - PASSWORD_ACCESS=***
  #   volumes:
  #     - /mnt/HoardingCow_docker_data/openclaw/ssh-config:/config
  #     - /home/gortium/infra:/data/workspace/infra:ro
--- a/ai/hermes/Dockerfile
+++ b/ai/hermes/Dockerfile
@@ -20,10 +20,16 @@ RUN --mount=type=ssh \
    GIT_SSH_COMMAND='ssh -p 2222 -o StrictHostKeyChecking=no' \
    git clone --depth 1 --branch main \
    git@code.lazyworkhorse.net:gortium/hermes-agent.git fork && \
-    rm -rf fork/node_modules fork/.venv fork/.git && \
-    cp -a fork/. /opt/hermes/ && \
+    rsync -a --delete fork/ /opt/hermes/ \
+      --exclude node_modules \
+      --exclude .venv \
+      --exclude .git && \
    rm -rf /tmp/fork /root/.ssh/

+# ---------- Rebuild web UI ----------
+# Source files changed; node_modules (from base image) reused.
+RUN cd /opt/hermes && npm run build
+
 # ---------- Reinstall Python package (editable) ----------
 # Picks up source changes from our fork.
 RUN . /opt/hermes/.venv/bin/activate && \
@@ -34,7 +40,6 @@ USER root
 RUN apt-get update && \
    apt-get install -y --no-install-recommends \
        libportaudio2 ca-certificates poppler-utils imagemagick \
-        libolm-dev \
        texlive-latex-base texlive-latex-extra texlive-fonts-recommended \
        texlive-xetex texlive-science \
        qemu-user-static binfmt-support emacs-nox && \
@@ -43,20 +48,6 @@ RUN apt-get update && \
 # ---------- UV ----------
 COPY --chmod=0755 --from=ghcr.io/astral-sh/uv:latest /uv /usr/local/bin/

-# ---------- Matrix bridge + extra pip deps ----------
-# Previously installed inline at container startup and persisted via volume mount.
-# Now baked into the image so the fragile venv volume mount can be removed.
-RUN . /opt/hermes/.venv/bin/activate && \
-    uv pip install --no-cache-dir 'mautrix[encryption]' openai
-
-WORKDIR /opt/hermes
-
-# ---------- Matrix bridge + extra pip deps ----------
-# Previously installed inline at container startup and persisted via volume mount.
-# Now baked into the image so the fragile venv volume mount can be removed.
-RUN . /opt/hermes/.venv/bin/activate && \
-    uv pip install --no-cache-dir 'mautrix[encryption]' openai
-
 # ---------- Piper TTS ----------
 RUN . /opt/hermes/.venv/bin/activate && \
    uv pip install --no-cache-dir piper-tts sounddevice numpy && \
@@ -84,9 +75,29 @@ os.remove(tgz)
 print('himalaya v1.2.0 installed')
 PYEOF

-# ---------- Install multi-gateway launcher ----------
-# Launches one gateway process per profile (HERMES_PROFILES env var)
-COPY --chmod=0755 run-multi-gateways.sh /usr/local/bin/run-multi-gateways.sh
+# ---------- Install himalaya-ro wrapper ----------
+COPY --chmod=0755 himalaya-ro.sh /usr/local/bin/himalaya-ro
+
+
+# ---------- Install 7-Zip (7zz) for CHM extraction ----------
+RUN /opt/hermes/.venv/bin/python3 /dev/stdin << 'PYEOF'
+import urllib.request, tarfile, os, shutil, subprocess
+url = 'https://github.com/ip7z/7zip/releases/download/26.01/7z2601-linux-x64.tar.xz'
+xz = '/tmp/7z2601-linux-x64.tar.xz'
+urllib.request.urlretrieve(url, xz)
+os.makedirs('/tmp/7z', exist_ok=True)
+with tarfile.open(xz, 'r:xz') as t:
+    t.extractall('/tmp/7z')
+shutil.move('/tmp/7z/7zz', '/usr/local/bin/7zz')
+os.chmod('/usr/local/bin/7zz', 0o755)
+shutil.rmtree('/tmp/7z', ignore_errors=True)
+os.remove(xz)
+# Verify
+result = subprocess.run(['/usr/local/bin/7zz'], capture_output=True, text=True)
+assert result.returncode == 0, f'7zz verify failed: {result.stderr}'
+print('7-Zip 26.01 installed successfully')
+PYEOF
+

 # ---------- Runtime ----------
 USER hermes
@@ -97,7 +108,6 @@ ENV CHROME_EXECUTABLE=/opt/hermes/.playwright/chromium/chrome-linux/chrome

 # Ensure tools directory and toolsets.py are writable by the hermes runtime user
 # so custom tools can be injected from the persistent volume at startup.
-USER root
 RUN chown -R hermes:hermes /opt/hermes/tools /opt/hermes/toolsets.py

 VOLUME [ "/opt/data" ]
--- a/ai/hermes/himalaya-ro.sh
+++ b/ai/hermes/himalaya-ro.sh
@@ -0,0 +1,73 @@
+#!/usr/bin/env bash
+# ─────────────────────────────────────────────────────────────
+# himalaya-ro — Read-only wrapper for himalaya
+#
+# Blocks destructive commands and logs audit trail.
+# Pass-through for read-only commands (list, read, search).
+#
+# Usage:  himalaya-ro [options] <command> [args...]
+#
+# Install: place in PATH before the real himalaya, or use
+#          `ln -sf himalaya-ro /usr/local/bin/himalaya`
+# ─────────────────────────────────────────────────────────────
+set -o pipefail
+
+# ── Configuration ───────────────────────────────────────────
+HIMALAYA_BIN="${HIMALAYA_BIN:-/usr/local/bin/himalaya}"
+AUDIT_LOG="${HIMALAYA_AUDIT_LOG:-/var/log/himalaya-audit.log}"
+
+# ── Destructive commands we block ──────────────────────────
+BLOCKED_CMDS=(
+  "message move"
+  "message delete"
+  "message copy"
+  "flag add"
+  "flag remove"
+  "folder create"
+  "folder delete"
+  "folder rename"
+  "template send"
+  "account configure"
+  "account delete"
+)
+
+# ── Determine the subcommand being invoked ─────────────────
+# Strip leading options (--account, --output, etc.) to find the verb
+ARGS=()
+SKIP_NEXT=false
+for arg in "$@"; do
+  if $SKIP_NEXT; then
+    SKIP_NEXT=false
+    continue
+  fi
+  if [[ "$arg" == --* ]]; then
+    case "$arg" in
+      --account|--output|--page|--page-size|--folder|--color|--format)
+        SKIP_NEXT=true ;;
+    esac
+    continue
+  fi
+  ARGS+=("$arg")
+done
+
+# Build subcommand string and check against blocklist
+CMD_STR=""
+for ((i=0; i<${#ARGS[@]}; i++)); do
+  if [ -z "$CMD_STR" ]; then
+    CMD_STR="${ARGS[$i]}"
+  else
+    CMD_STR="$CMD_STR ${ARGS[$i]}"
+  fi
+  for blocked in "${BLOCKED_CMDS[@]}"; do
+    if [[ "$CMD_STR" == "$blocked" ]]; then
+      TS=$(date '+%Y-%m-%d %H:%M:%S')
+      echo "[AUDIT] $TS BLOCKED: himalaya $*" >> "$AUDIT_LOG"
+      echo "ERROR: Command 'himalaya $CMD_STR ...' is blocked by read-only policy." >&2
+      echo "       Audit log: $AUDIT_LOG" >&2
+      exit 100
+    fi
+  done
+done
+
+# ── Allow pass-through ─────────────────────────────────────
+exec "$HIMALAYA_BIN" "$@"
--- a/ai/hermes/run-multi-gateways.sh
+++ b/ai/hermes/run-multi-gateways.sh
@@ -1,32 +0,0 @@
-#!/bin/bash
-# Multi-gateway launcher for HERMES_PROFILES env var.
-# Reads comma-separated profile names, spawns one gateway per profile.
-# Designed to run before the main entrypoint — gateways run in background.
-set -e
-
-if [ -z "${HERMES_PROFILES}" ]; then
-  echo "HERMES_PROFILES not set — skipping multi-gateway launch"
-  exit 0
-fi
-
-# Source venv to make 'hermes' available (entrypoint.sh sources it later,
-# but we need it NOW for the background gateways)
-HERMES_BIN="/opt/hermes/.venv/bin/hermes"
-if [ ! -x "$HERMES_BIN" ]; then
-  echo "ERROR: hermes binary not found at $HERMES_BIN"
-  exit 1
-fi
-
-mkdir -p /opt/data/logs
-
-IFS=',' read -ra PROFILES <<< "${HERMES_PROFILES}"
-for profile in "${PROFILES[@]}"; do
-  profile="$(echo "${profile}" | xargs)"  # trim whitespace
-  [ -z "${profile}" ] && continue
-
-  echo "Starting gateway for profile: ${profile}"
-  nohup env API_SERVER_ENABLED=false API_SERVER_KEY= gosu hermes "$HERMES_BIN" --profile "${profile}" gateway run \
-      >> "/opt/data/logs/gateway-${profile}.log" 2>&1 &
-done
-
-echo "All gateways launched: ${HERMES_PROFILES}"
--- a/ai/honcho/Dockerfile
+++ b/ai/honcho/Dockerfile
@@ -1,75 +1,72 @@
-# build stage — fetches and builds Honcho from source
-FROM python:3.13-slim-bookworm AS honcho-builder
+# Honcho — Memory infrastructure for stateful AI agents
+# Builds the Honcho FastAPI server from the official GitHub repository.
+#
+# Usage:
+#   docker compose build honcho
+#   docker compose up honcho
+#
+# Reference: https://github.com/plastic-labs/honcho

-RUN apt-get update && \
-    apt-get install -y --no-install-recommends git openssh-client && \
-    rm -rf /var/lib/apt/lists/*
+# ---------------------------------------------------------------------------
+# Stage 1 — clone source & install dependencies
+# ---------------------------------------------------------------------------
+FROM python:3.13-slim-bookworm AS builder
+
+RUN apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/*

 COPY --from=ghcr.io/astral-sh/uv:0.9.24 /uv /bin/uv

-ARG HONCHO_REPO=ssh://git@code.lazyworkhorse.net:2222/Hermes/honcho.git
-ARG HONCHO_REF=main
-RUN mkdir -p -m 0700 ~/.ssh && ssh-keyscan -p 2222 code.lazyworkhorse.net >> ~/.ssh/known_hosts 2>/dev/null
-RUN --mount=type=ssh git clone --depth 1 --branch ${HONCHO_REF} ${HONCHO_REPO} /app
+WORKDIR /src
+RUN git clone --depth 1 --branch main https://github.com/plastic-labs/honcho.git .
+
+ENV UV_COMPILE_BYTECODE=1
+ENV UV_LINK_MODE=copy
+
+# Install project dependencies (frozen from lockfile, no dev)
+RUN --mount=type=cache,target=/root/.cache/uv \
+    uv sync --frozen --no-install-project --no-group dev
+
+# ---------------------------------------------------------------------------
+# Stage 2 — runtime image
+# ---------------------------------------------------------------------------
+FROM python:3.13-slim-bookworm AS runtime
+
+COPY --from=ghcr.io/astral-sh/uv:0.9.24 /uv /bin/uv
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ca-certificates && \
+    rm -rf /var/lib/apt/lists/*

 WORKDIR /app

 ENV UV_COMPILE_BYTECODE=1
 ENV UV_LINK_MODE=copy
-ENV UV_PYTHON=/usr/local/bin/python3.13
-
-RUN uv sync --frozen
-
-# build stage — builds OpenConcho SPA
-FROM node:22-bookworm AS openconcho-builder
-
-ENV PNPM_HOME=/pnpm
-ENV PATH=$PNPM_HOME:$PATH
-RUN corepack enable && corepack prepare pnpm@latest --activate
-
-WORKDIR /app
-RUN apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/*
-
-ARG OPENCONCHO_SHA=e490d911fcb27ee193558fd9a28856cde2057665
-RUN git clone --depth 1 https://github.com/offendingcommit/openconcho.git /app && \
-    git -C /app fetch --depth 1 origin ${OPENCONCHO_SHA} && \
-    git -C /app checkout ${OPENCONCHO_SHA}
-
-RUN pnpm install --frozen-lockfile
-RUN pnpm --filter @openconcho/web build
-
-# runtime stage — nginx + Honcho FastAPI
-FROM python:3.13-slim-bookworm
-
-# Install nginx and create runtime dirs before dropping permissions
-RUN apt-get update && apt-get install -y --no-install-recommends nginx && \
-    rm -rf /var/log/nginx/* && \
-    rm -rf /var/lib/apt/lists/* && \
-    rm -f /etc/nginx/sites-enabled/default
-
-# Patch nginx.conf: comment out "user www-data;" so nginx master stays as root
-# (workers inherit root inside a container — fine for single-service isolation)
-RUN sed -i 's/^user /# user /' /etc/nginx/nginx.conf
-
-# Pre-create nginx runtime directories with proper ownership
-RUN mkdir -p /var/lib/nginx/body /var/lib/nginx/proxy /var/lib/nginx/fastcgi \
-             /var/lib/nginx/uwsgi /var/lib/nginx/scgi /var/lib/nginx/proxy_temp \
-             /var/cache/nginx && \
-    chown -R root:root /var/lib/nginx /var/cache/nginx
-
-# Honcho
-COPY --from=honcho-builder /app /app
-WORKDIR /app
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
 ENV PATH="/app/.venv/bin:$PATH"
 ENV HOME=/app
-COPY config.toml /app/config.toml
+ENV UV_CACHE_DIR=/tmp/uv-cache

-# OpenConcho SPA
-COPY --from=openconcho-builder /app/packages/web/dist /usr/share/nginx/html
+# Copy the dependency layer from the builder
+COPY --from=builder /src/uv.lock /src/pyproject.toml /app/
+RUN --mount=type=cache,target=/root/.cache/uv \
+    uv sync --frozen --no-group dev

-# nginx config (proxies /v3/, /v2/ to Honcho on localhost:8000)
-COPY honcho-nginx.conf /etc/nginx/conf.d/default.conf
+# Copy application source and config
+COPY --from=builder /src/src/ /app/src/
+COPY --from=builder /src/migrations/ /app/migrations/
+COPY --from=builder /src/scripts/ /app/scripts/
+COPY --from=builder /src/docker/ /app/docker/
+COPY --from=builder /src/alembic.ini /app/alembic.ini

-EXPOSE 80
+# Create non-root user
+RUN addgroup --system app && \
+    adduser --system --ingroup app app && \
+    mkdir -p /tmp/uv-cache && \
+    chown -R app:app /app /tmp/uv-cache

-CMD ["bash", "-c", "nginx -g 'daemon off;' & fastapi run --host 127.0.0.1 --port 8000 src/main.py & python3 -m src.deriver & wait -n"]
+USER app
+EXPOSE 8000
+
+# The entrypoint.sh script runs database migrations then starts the FastAPI server
+ENTRYPOINT ["sh", "docker/entrypoint.sh"]
--- a/ai/honcho/config.toml
+++ b/ai/honcho/config.toml
@@ -1,132 +0,0 @@
-[app]
-LOG_LEVEL = "INFO"
-MAX_MESSAGE_SIZE = 25000
-EMBED_MESSAGES = true
-NAMESPACE = "honcho"
-
-[db]
-CONNECTION_URI = "postgresql+psycopg://honcho:honcho_pass@honcho-db:5432/honcho"
-SCHEMA = "public"
-POOL_SIZE = 10
-MAX_OVERFLOW = 20
-
-[auth]
-USE_AUTH = false
-
-[sentry]
-ENABLED = false
-
-[telemetry]
-ENABLED = false
-
-[webhook]
-ENABLED = false
-
-[cache]
-ENABLED = true
-URL = "redis://honcho-redis:6379/0"
-
-[llm]
-DEFAULT_MAX_TOKENS = 4096
-
-# Embeddings via Ollama — bge-m3 provides 1024-dim
-[embedding]
-VECTOR_DIMENSIONS = 1024
-MAX_INPUT_TOKENS = 8192
-
-[embedding.model_config]
-transport = "openai"
-model = "bge-m3"
-overrides = {base_url = "http://ollama:11434/v1", api_key = "ollama"}
-
-# --- Deriver ---
-[deriver]
-ENABLED = true
-WORKERS = 1
-POLLING_SLEEP_INTERVAL_SECONDS = 5.0
-FLUSH_ENABLED = true
-
-[deriver.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-# --- Dialectic ---
-[dialectic]
-MAX_INPUT_TOKENS = 4096
-SESSION_HISTORY_MAX_TOKENS = 8192
-
-[dialectic.levels.minimal]
-MAX_TOOL_ITERATIONS = 1
-MAX_OUTPUT_TOKENS = 512
-[dialectic.levels.minimal.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-[dialectic.levels.low]
-MAX_TOOL_ITERATIONS = 3
-[dialectic.levels.low.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-[dialectic.levels.medium]
-MAX_TOOL_ITERATIONS = 2
-[dialectic.levels.medium.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-[dialectic.levels.high]
-MAX_TOOL_ITERATIONS = 4
-[dialectic.levels.high.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-[dialectic.levels.max]
-MAX_TOOL_ITERATIONS = 10
-[dialectic.levels.max.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-# --- Summary ---
-[summary]
-ENABLED = true
-MESSAGES_PER_SHORT_SUMMARY = 20
-MESSAGES_PER_LONG_SUMMARY = 60
-
-[summary.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-# --- Dream ---
-[dream]
-ENABLED = true
-
-[dream.model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-[dream.deduction_model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-[dream.induction_model_config]
-overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
-transport = "openai"
-model = "deepseek-v4-flash"
-
-# --- Peer Card ---
-[peer_card]
-ENABLED = true
-
-# --- Vector Store ---
-[vector_store]
-TYPE = "pgvector"
-# DIMENSIONS is deprecated — EMBEDDING.VECTOR_DIMENSIONS is authoritative
--- a/ai/honcho/honcho-nginx.conf
+++ b/ai/honcho/honcho-nginx.conf
@@ -1,52 +0,0 @@
-server {
-    listen 80 default_server;
-    listen [::]:80 default_server;
-    server_name _;
-
-    root /usr/share/nginx/html;
-    index index.html;
-
-    # Honcho API proxy
-    location /v3/ {
-        proxy_pass http://127.0.0.1:8000;
-        proxy_http_version 1.1;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-    }
-
-    location /v2/ {
-        proxy_pass http://127.0.0.1:8000;
-        proxy_http_version 1.1;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-    }
-
-    # Honcho health
-    location /health {
-        proxy_pass http://127.0.0.1:8000;
-        proxy_http_version 1.1;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-    }
-
-    # OpenAPI docs
-    location /openapi.json {
-        proxy_pass http://127.0.0.1:8000;
-        proxy_http_version 1.1;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-    }
-
-    # SPA: fallback to index.html for client-side routing
-    location / {
-        try_files $uri $uri/ /index.html;
-    }
-}
--- a/ai/honcho/init-db.sql
+++ b/ai/honcho/init-db.sql
--- a/build/honcho/config.toml
+++ b/build/honcho/config.toml
@@ -0,0 +1,93 @@
+# Honcho Configuration
+# Pre-configured for self-hosted deployment with Ollama embeddings.
+# Mount this file at /app/config.toml in the Honcho container.
+#
+# Environment variables override these values at runtime
+# (e.g. DB_CONNECTION_URI, DERIVER_*).
+
+[app]
+LOG_LEVEL = "INFO"
+NAMESPACE = "honcho"
+SESSION_OBSERVERS_LIMIT = 10
+GET_CONTEXT_MAX_TOKENS = 16384
+EMBED_MESSAGES = true
+
+[db]
+# Connection URI is set via environment variable DB_CONNECTION_URI
+SCHEMA = "public"
+POOL_SIZE = 10
+MAX_OVERFLOW = 20
+POOL_TIMEOUT = 30
+POOL_RECYCLE = 300
+POOL_PRE_PING = true
+POOL_USE_LIFO = true
+SQL_DEBUG = false
+
+[auth]
+USE_AUTH = false
+
+[llm]
+DEFAULT_MAX_TOKENS = 4096
+
+[embedding]
+VECTOR_DIMENSIONS = 768
+MAX_INPUT_TOKENS = 8192
+MAX_TOKENS_PER_REQUEST = 2048
+
+[embedding.model_config]
+transport = "openai"
+model = "nomic-embed-text:latest"
+
+[embedding.model_config.overrides]
+base_url = "http://ollama:11434/v1"
+# Ollama does not require an API key; env var must be set to non-empty string
+api_key_env = "LLM_OPENAI_API_KEY"
+
+[deriver]
+ENABLED = false
+WORKERS = 1
+POLLING_SLEEP_INTERVAL_SECONDS = 1.0
+STALE_SESSION_TIMEOUT_MINUTES = 5
+DEDUPLICATE = true
+LOG_OBSERVATIONS = false
+
+[deriver.model_config]
+transport = "openai"
+model = "qwen3.6:27b-q4_K_M"
+
+[deriver.model_config.overrides]
+base_url = "http://ollama:11434/v1"
+api_key_env = "LLM_OPENAI_API_KEY"
+
+[summary]
+ENABLED = false
+
+[summary.model_config]
+transport = "openai"
+model = "qwen3.6:27b-q4_K_M"
+
+[summary.model_config.overrides]
+base_url = "http://ollama:11434/v1"
+api_key_env = "LLM_OPENAI_API_KEY"
+
+[dream]
+ENABLED = false
+
+[dialectic]
+MAX_OUTPUT_TOKENS = 4096
+MAX_INPUT_TOKENS = 16384
+
+[cache]
+ENABLED = false
+
+[vector_store]
+TYPE = "pgvector"
+
+[metrics]
+ENABLED = false
+
+[telemetry]
+ENABLED = false
+
+[sentry]
+ENABLED = false
--- a/copy_script.txt
+++ b/copy_script.txt
@@ -1,7 +0,0 @@
-import shutil, os
-src = "/opt/data/projects/gortium/compose/ai/compose_updated.txt"
-dst = "/opt/data/projects/gortium/compose/ai/compose.yml"
-print(f"Source exists: {os.path.exists(src)}, size: {os.path.getsize(src)}")
-print(f"Dest exists: {os.path.exists(dst)}")
-shutil.copy2(src, dst)
-print(f"Copied, dest size: {os.path.getsize(dst)}")
--- a/entrypoint-combined.sh
+++ b/entrypoint-combined.sh
@@ -1,154 +0,0 @@
-#!/bin/bash
-# ── Hermes Workspace Combined Entrypoint ──
-# Waits for the Hermes gateway container (hermes:8642) to become healthy,
-# then starts the Hermes Workspace web UI in the foreground.
-# Supports graceful shutdown via SIGTERM/SIGINT.
-# ──────────────────────────────────────────
-
-set -euo pipefail
-
-# ── Configuration ──────────────────────────────────────────────
-GATEWAY_HOST="${GATEWAY_HOST:-hermes}"
-GATEWAY_PORT="${GATEWAY_PORT:-8642}"
-GATEWAY_URL="http://${GATEWAY_HOST}:${GATEWAY_PORT}"
-
-HEALTH_ENDPOINT="${HEALTH_ENDPOINT:-/health}"
-MAX_RETRIES="${HEALTH_MAX_RETRIES:-60}"
-RETRY_INTERVAL="${HEALTH_RETRY_INTERVAL:-2}"
-
-WORKSPACE_DIR="${WORKSPACE_DIR:-/workspace}"
-WORKSPACE_ENTRY="${WORKSPACE_ENTRY:-server-entry.js}"
-
-PID_FILE="${PID_FILE:-/tmp/workspace.pid}"
-
-# ── Logging ────────────────────────────────────────────────────
-log_info()  { echo "[$(date '+%Y-%m-%d %H:%M:%S')] [INFO]  $*"; }
-log_warn()  { echo "[$(date '+%Y-%m-%d %H:%M:%S')] [WARN]  $*"; }
-log_error() { echo "[$(date '+%Y-%m-%d %H:%M:%S')] [ERROR] $*"; }
-
-# ── Graceful Shutdown ──────────────────────────────────────────
-_workspace_pid=""
-_shutting_down=false
-
-cleanup() {
-    if [ "$_shutting_down" = true ]; then
-        return
-    fi
-    _shutting_down=true
-
-    log_info "Shutdown signal received, cleaning up..."
-
-    # Stop workspace process if running
-    if [ -n "$_workspace_pid" ] && kill -0 "$_workspace_pid" 2>/dev/null; then
-        log_info "Stopping workspace (PID: $_workspace_pid)..."
-        kill -TERM "$_workspace_pid" 2>/dev/null || true
-
-        # Give it time to shut down gracefully
-        local wait_sec=10
-        while kill -0 "$_workspace_pid" 2>/dev/null && [ "$wait_sec" -gt 0 ]; do
-            sleep 1
-            wait_sec=$((wait_sec - 1))
-        done
-
-        # Force kill if still running
-        if kill -0 "$_workspace_pid" 2>/dev/null; then
-            log_warn "Workspace did not shut down gracefully, force killing..."
-            kill -KILL "$_workspace_pid" 2>/dev/null || true
-        fi
-    fi
-
-    # Clean up PID file
-    [ -f "$PID_FILE" ] && rm -f "$PID_FILE"
-
-    log_info "Shutdown complete."
-    exit 0
-}
-
-# Trap termination signals for graceful shutdown
-trap cleanup SIGTERM SIGINT
-
-# ── Gateway Health Check ───────────────────────────────────────
-wait_for_gateway() {
-    local url="${GATEWAY_URL}${HEALTH_ENDPOINT}"
-    local retries="$MAX_RETRIES"
-    local interval="$RETRY_INTERVAL"
-    local attempt=0
-
-    log_info "Waiting for Hermes gateway at ${GATEWAY_URL}..."
-    log_info "Max retries: ${retries}, interval: ${interval}s"
-
-    while [ "$attempt" -lt "$retries" ]; do
-        attempt=$((attempt + 1))
-
-        if curl -fsS "${url}" >/dev/null 2>&1; then
-            log_info "Gateway is healthy after ${attempt} attempt(s) (${GATEWAY_URL})"
-            return 0
-        fi
-
-        if [ "$attempt" -lt "$retries" ]; then
-            log_info "Gateway not ready yet (attempt ${attempt}/${retries}), retrying in ${interval}s..."
-            sleep "$interval"
-        fi
-    done
-
-    log_error "Gateway did not become healthy after ${retries} attempts (${retries * interval}s)"
-    return 1
-}
-
-# ── Workspace Startup ──────────────────────────────────────────
-start_workspace() {
-    local entry="${WORKSPACE_DIR}/${WORKSPACE_ENTRY}"
-
-    if [ ! -d "$WORKSPACE_DIR" ]; then
-        log_error "Workspace directory not found: ${WORKSPACE_DIR}"
-        return 1
-    fi
-
-    if [ ! -f "$entry" ]; then
-        log_error "Workspace entry point not found: ${entry}"
-        return 1
-    fi
-
-    log_info "Starting Hermes Workspace web UI..."
-    log_info "  Directory: ${WORKSPACE_DIR}"
-    log_info "  Entry:     ${entry}"
-
-    cd "$WORKSPACE_DIR"
-
-    # Start workspace in background so we can trap signals
-    exec node --max-old-space-size=2048 "${entry}" &
-    _workspace_pid=$!
-    echo "$_workspace_pid" > "$PID_FILE"
-
-    log_info "Workspace started (PID: ${_workspace_pid})"
-
-    # Wait for workspace process
-    wait "$_workspace_pid"
-    local exit_code=$?
-
-    log_info "Workspace exited with code ${exit_code}"
-    return "$exit_code"
-}
-
-# ── Main ───────────────────────────────────────────────────────
-main() {
-    log_info "=== Hermes Workspace Combined Entrypoint ==="
-    log_info "Gateway:   ${GATEWAY_URL}"
-    log_info "Workspace: ${WORKSPACE_DIR}/${WORKSPACE_ENTRY}"
-    log_info "PID file:  ${PID_FILE}"
-
-    # Wait for gateway to be healthy
-    if ! wait_for_gateway; then
-        log_warn "Proceeding without confirmed gateway health..."
-    fi
-
-    # Start the workspace
-    start_workspace
-    local exit_code=$?
-
-    log_info "Entrypoint exiting with code ${exit_code}"
-    return "$exit_code"
-}
-
-# Run main; exit with its return code
-main "$@"
--- a/env/.env.example.honcho
+++ b/env/.env.example.honcho
@@ -0,0 +1,31 @@
+# Honcho Environment Variables
+# Copy this file to your .env (at the compose root or docker-compose working directory)
+# and fill in the secrets.
+#
+#   cp env/.env.example.honcho .env
+#
+# Then reference it from compose.yml:
+#   env_file:
+#     - path: .env
+#       required: true
+
+# ---------------------------------------------------------------------------
+# Database
+# ---------------------------------------------------------------------------
+# PostgreSQL connection string for Honcho.
+# The password must match HONCHO_DB_PASSWORD below.
+HONCHO_DB_PASSWORD=change_me_to_a_strong_random_password
+
+# ---------------------------------------------------------------------------
+# LLM Provider
+# ---------------------------------------------------------------------------
+# Ollama does not require a real API key, but the env var must be set to a
+# non-empty string for the OpenAI-compatible client to connect.
+LLM_OPENAI_API_KEY=ollama
+
+# ---------------------------------------------------------------------------
+# Honcho Server
+# ---------------------------------------------------------------------------
+# Honcho will pick up DB_CONNECTION_URI from the compose environment.
+# You can override additional settings here if needed.
+# LOG_LEVEL=INFO
--- a/replace_compose.py
+++ b/replace_compose.py
@@ -1,17 +0,0 @@
-#!/usr/bin/env python3
-"""Copy the updated txt file over the original yml file."""
-import shutil
-import os
-
-src = "/opt/data/projects/gortium/compose/ai/compose_updated.txt"
-dst = "/opt/data/projects/gortium/compose/ai/compose.yml"
-
-# Check src exists
-print(f"Source exists: {os.path.exists(src)}")
-print(f"Source size: {os.path.getsize(src)} bytes")
-print(f"Destination exists: {os.path.exists(dst)}")
-
-# Copy
-shutil.copy2(src, dst)
-print(f"Copied {src} -> {dst}")
-print(f"Destination size: {os.path.getsize(dst)} bytes")