Compare commits

..

26 Commits

Author SHA1 Message Date
5765dc6004 Merge remote-tracking branch 'origin/fix/hermes-build' 2026-05-22 21:13:46 -04:00
e8075fb71b fix: also clear API_SERVER_KEY for profile gateways
Line 1521 in gateway/config.py: if api_server_enabled or api_server_key:
The compose.yml sets API_SERVER_KEY=hermes_local_key, which was enough
to enable the API server even with API_SERVER_ENABLED=false.
2026-05-22 13:53:53 -04:00
645d519030 fix: use env to force API_SERVER_ENABLED=false in multi-gateway launcher
Shell prefix didn't work with nohup+gosu chain - Docker compose
env var API_SERVER_ENABLED=true leaked through. Using 'env'
command guarantees the override is in the child process env.
2026-05-22 13:52:05 -04:00
146add2a64 fix: use full hermes path and gosu in multi-gateway launcher
- Use /opt/hermes/.venv/bin/hermes (full path) — not on PATH
  before entrypoint.sh sources the venv
- Wrap with gosu hermes to avoid root guard in gateway run
- Add error check if hermes binary doesn't exist
2026-05-22 13:34:12 -04:00
eb3795d9e3 Merge remote-tracking branch 'origin/fix/hermes-build' 2026-05-22 13:26:59 -04:00
ef6e0e57b6 Merge remote-tracking branch 'origin/fix/remove-venv-mount' 2026-05-22 13:25:59 -04:00
d8a1ebcd96 fix: add missing command: gateway run to hermes service
Without this,  is empty and entrypoint.sh runs bare 'hermes'
which defaults to interactive chat mode. With a non-TTY stdin
this exits immediately with prompt_toolkit's 'Input is not a
terminal' warning, causing a container restart loop.

The profile gateways (run-multi-gateways.sh) were unaffected
because the script passes 'gateway run' explicitly.
2026-05-22 13:23:08 -04:00
45a224eb7f fix: add missing command: gateway run to hermes service
Without this,  is empty and entrypoint.sh runs bare 'hermes'
which defaults to interactive chat mode. With a non-TTY stdin
this exits immediately with prompt_toolkit's 'Input is not a
terminal' warning, causing a container restart loop.

The profile gateways (run-multi-gateways.sh) were unaffected
because the script passes 'gateway run' explicitly.
2026-05-22 13:22:58 -04:00
46f3b23a3f Merge pull request 'fix: remove venv volume mount conflicting with entrypoint.sh' (#49) from fix/remove-venv-mount into master
Some checks failed
Build Hermes agent / build (push) Has been cancelled
Build ollama (gfx906) / build (push) Has been cancelled
Reviewed-on: #49
2026-05-22 17:05:52 +00:00
567850bd13 feat: bake Matrix bridge deps into Docker image instead of volume mount
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
- Add libolm-dev system dep (required by mautrix[encryption])
- Add mautrix[encryption] + openai pip packages to build
- These were previously installed inline at container startup and
  persisted via the fragile venv volume mount (now removed)
2026-05-22 13:04:49 -04:00
bce336c4fd feat: bake Matrix bridge deps into Docker image instead of volume mount
- Add libolm-dev system dep (required by mautrix[encryption])
- Add mautrix[encryption] + openai pip packages to build
- These were previously installed inline at container startup and
  persisted via the fragile venv volume mount (now removed)
2026-05-22 13:04:43 -04:00
e4117cd3d5 fix: remove venv volume mount conflicting with entrypoint.sh
The volume mount at /mnt/HoardingCow_docker_data/Hermes/venv overrides the
container's built .venv with an empty or stale host directory, causing
entrypoint.sh line 62 to fail on 'source .venv/bin/activate' (set -e).

The Docker image already builds a complete venv — no need to persist it.
2026-05-22 13:04:08 -04:00
2d59bb44c3 fix: remove venv volume mount conflicting with upstream entrypoint.sh
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
The volume mount /mnt/HoardingCow_docker_data/Hermes/venv overrides the
container's built-in .venv with whatever is on the host. On a fresh start
or after a clean build, an empty/missing venv directory causes entrypoint.sh
line 62 (source .venv/bin/activate) to fail with set -e.

The Docker image already builds a complete venv — persisting it on the host
is unnecessary and fragile.
2026-05-22 13:03:51 -04:00
bf56d4be8b Merge pull request 'fix: resolve Docker build errors and add Traefik routing for Hermes web UI' (#33) from fix/hermes-build into master
Some checks failed
Build Hermes agent / build (push) Has been cancelled
Build ollama (gfx906) / build (push) Has been cancelled
Reviewed-on: #33
2026-05-22 16:36:23 +00:00
1c8efb1090 Merge remote-tracking branch 'origin/master' into fix/hermes-build
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
# Conflicts:
#	ai/compose.yml
2026-05-22 12:34:39 -04:00
7725830e6c feat: wire up HERMES_PROFILES to multi-gateway launcher script
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
- Add run-multi-gateways.sh to /opt/data/hermes-tools/ that reads
  HERMES_PROFILES env var and spawns one gateway per profile
- Update entrypoint to call the script before the main entrypoint
- Set HERMES_PROFILES=ashley,claire,finn,matt,paul (was default)

Closes PR #47 (feat/multi-profile-gateways). Builds on 548e15d's cleaner
env-var-driven approach — compose.yml stays declarative, logic in script.
2026-05-22 12:27:47 -04:00
548e15d6b4 feat(compose): add HERMES_PROFILES env var for multi-gateway mode
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
2026-05-20 20:02:26 -04:00
35aa466e87 fix: Matrix bridge ModuleNotFoundError - install deps to venv with persistence
Some checks failed
Build Hermes agent / build (push) Has been cancelled
Build ollama (gfx906) / build (push) Has been cancelled
fix: Matrix bridge ModuleNotFoundError - install deps to venv with persistence
2026-05-20 18:47:28 +00:00
2ff99b1f57 Merge master into fix/matrix-bridge-dependencies to resolve conflicts
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
2026-05-20 14:42:02 -04:00
dea1429a5d Merge pull request #2: fix: Matrix bridge ModuleNotFoundError - install deps to venv with persistence
# Conflicts:
#	ai/compose.yml
2026-05-20 14:41:06 -04:00
ebad994d60 feat(hermes): enable dashboard (HERMES_DASHBOARD=1) + Authelia auth
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
2026-05-20 14:06:23 -04:00
5f25c87775 fix: add missing USER root before chown step
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
The chown -R hermes:hermes was running as non-root user 'hermes'
since USER hermes was set earlier. The new upstream base image
(v0.12.0+) has tools/ owned by root, so the chown fails.
Previous base image happened to have tools/ owned by hermes,
making the chown a silent no-op.
2026-05-20 13:32:18 -04:00
4e566b2408 fix: resolve Docker build errors and add Traefik routing for Hermes web UI
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Build ollama (gfx906) / build (pull_request) Has been cancelled
- Replace rsync with cp -a (rsync unavailable in latest upstream base image)
- Remove npm run build step (fork's package.json has no build script)
- Remove himalaya-ro.sh from build context (deployed via install.sh)
- Add hermes to ai_net network for Traefik access
- Add Traefik labels routing hermes.lazyworkhorse.net to dashboard port 9119
2026-05-20 13:14:10 -04:00
Thierry Pouplier
2aab06cc1a fix: use mautrix[encryption] instead of matrix-nio for Matrix bridge
The Hermes Matrix gateway uses the mautrix SDK, not matrix-nio.
This fixes E2EE support by installing the correct library.

Refs: PR #2
2026-04-29 03:34:15 +00:00
Thierry Pouplier
a404f5e2c4 fix: Correct OPENROUTER_API_KEY variable name 2026-04-29 02:43:35 +00:00
Thierry Pouplier
f9afd79f3e fix: Add openai and matrix-nio dependencies for Hermes Matrix bridge 2026-04-29 02:19:24 +00:00
8 changed files with 76 additions and 451 deletions

183
ai/compose.yml Normal file → Executable file
View File

@@ -32,19 +32,24 @@ services:
- default
container_name: hermes
entrypoint: ["/bin/bash", "-c",
"bash /opt/data/hermes-tools/install.sh && exec /usr/bin/tini -g -- /opt/hermes/docker/entrypoint.sh \"$@\"",
"bash /opt/data/hermes-tools/install.sh && bash /opt/data/hermes-tools/run-multi-gateways.sh && exec /usr/bin/tini -g -- /opt/hermes/docker/entrypoint.sh \"$@\"",
"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=${OPEN...KEY}
- OPENROUTER_API_KEY=${OPENROUTER_API_KEY}
# ROCm for GPU-accelerated faster-whisper STT
- HSA_OVERRIDE_GFX_VERSION=9.0.6
- HCC_AMDGPU_TARGET=gfx906
@@ -66,6 +71,30 @@ services:
- "26"
networks:
- ai_backend
- ai_net
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
@@ -129,150 +158,6 @@ services:
- "303"
- "26"
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: ${HONCHO_DB_PASSWORD:?HONCHO_DB_PASSWORD must be set}
healthcheck:
test: ["CMD-SHELL", "pg_isready -U honcho -d honcho"]
interval: 5s
timeout: 5s
retries: 10
volumes:
- /mnt/HoardingCow_docker_data/Honcho/pgdata:/var/lib/postgresql/data
- ./honcho/init.sql:/docker-entrypoint-initdb.d/init.sql
networks:
- ai_backend
honcho:
build:
context: ./honcho
dockerfile: Dockerfile
container_name: honcho
restart: unless-stopped
ports:
- "127.0.0.1:8000:8000"
depends_on:
honcho-db:
condition: service_healthy
environment:
DB_CONNECTION_URI: postgresql+psycopg://honcho:${HONCHO_DB_PASSWORD:?HONCHO_DB_PASSWORD must be set}@honcho-db:5432/honcho
LOG_LEVEL: INFO
LLM_OPENAI_API_KEY: ${LLM_OPENAI_API_KEY:-ollama}
volumes:
- /mnt/HoardingCow_docker_data/Honcho/config.toml:/app/config.toml
networks:
- ai_backend
- ai_net
labels:
- "traefik.enable=true"
- "traefik.docker.network=ai_net"
- "traefik.http.routers.honcho-http.rule=Host(`honcho.lazyworkhorse.net`)"
- "traefik.http.routers.honcho-http.entrypoints=web"
- "traefik.http.routers.honcho-http.middlewares=redirect-to-https"
- "traefik.http.routers.honcho-https.rule=Host(`honcho.lazyworkhorse.net`)"
- "traefik.http.routers.honcho-https.entrypoints=websecure"
- "traefik.http.routers.honcho-https.tls=true"
- "traefik.http.routers.honcho-https.tls.certresolver=njalla"
- "traefik.http.services.honcho.loadbalancer.server.port=8000"
holographic-memory:
build:
context: ./holographic-memory
image: holographic-memory:latest
container_name: holographic-memory
restart: unless-stopped
ports:
- "127.0.0.1:8100:8100"
environment:
- HOLOGRAPHIC_DB_PATH=/data/holographic/memory_store.db
- HOLOGRAPHIC_PORT=8100
- HOLOGRAPHIC_DEFAULT_TRUST=0.5
volumes:
- /mnt/HoardingCow_docker_data/HolographicMemory:/data/holographic
networks:
- ai_backend
healthcheck:
test: ["CMD", "python3", "-c", "import urllib.request; urllib.request.urlopen('http://127.0.0.1:8100/health')"]
interval: 30s
timeout: 5s
retries: 3
start_period: 10s
networks:
ai_net:
external: true
@@ -424,8 +309,8 @@ networks:
# - /home/gortium/infra:/data/workspace/infra
# environment:
# - TZ=America/Toronto
# - OPENCLAW_GATEWAY_TOKEN=${OPEN...KEN}
# - OPENROUTER_API_KEY=${OPEN...KEY}
# - OPENCLAW_GATEWAY_TOKEN=${OPENCLAW_GATEWAY_TOKEN}
# - OPENROUTER_API_KEY=${OPENROUTER_API_KEY}
# # Point to the sidecar browser
# - BROWSER_CDP_URL=http://openclaw-browser:9222
# - BROWSER_EVALUATE_ENABLED=true
@@ -470,7 +355,7 @@ networks:
# - PGID=1000
# - PUBLIC_KEY_FILE=/config/ssh/authorized_keys
# - SUDO_ACCESS=false
# - PASSWORD_ACCESS=***
# - PASSWORD_ACCESS=false
# volumes:
# - /mnt/HoardingCow_docker_data/openclaw/ssh-config:/config
# - /home/gortium/infra:/data/workspace/infra:ro

View File

@@ -20,16 +20,10 @@ 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 && \
rsync -a --delete fork/ /opt/hermes/ \
--exclude node_modules \
--exclude .venv \
--exclude .git && \
rm -rf fork/node_modules fork/.venv fork/.git && \
cp -a fork/. /opt/hermes/ && \
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 && \
@@ -40,6 +34,7 @@ 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 && \
@@ -48,6 +43,12 @@ 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
# ---------- Piper TTS ----------
RUN . /opt/hermes/.venv/bin/activate && \
uv pip install --no-cache-dir piper-tts sounddevice numpy && \
@@ -75,30 +76,6 @@ os.remove(tgz)
print('himalaya v1.2.0 installed')
PYEOF
# ---------- 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
ENV HERMES_HOME=/opt/data
@@ -108,6 +85,7 @@ 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" ]

View File

@@ -1,73 +0,0 @@
#!/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" "$@"

View File

@@ -1,72 +0,0 @@
# 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
# ---------------------------------------------------------------------------
# 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
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 PYTHONDONTWRITEBYTECODE=1
ENV PYTHONUNBUFFERED=1
ENV PATH="/app/.venv/bin:$PATH"
ENV HOME=/app
ENV UV_CACHE_DIR=/tmp/uv-cache
# 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
# 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
# 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
USER app
EXPOSE 8000
# The entrypoint.sh script runs database migrations then starts the FastAPI server
ENTRYPOINT ["sh", "docker/entrypoint.sh"]

View File

@@ -1 +0,0 @@
CREATE EXTENSION IF NOT EXISTS vector;

View File

@@ -1,93 +0,0 @@
# 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

View File

@@ -1,31 +0,0 @@
# 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

32
run-multi-gateways.sh Executable file
View File

@@ -0,0 +1,32 @@
#!/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}"