code-whisperer

Docker Compose setup for running moltenai/moltenhub-code:vnext with a local faster-whisper speech service.

The stack starts:

• agent: runs harness hub --ui-listen :7777 and exposes the hub UI on http://localhost:3331.
• whisper: runs lscr.io/linuxserver/faster-whisper:latest, persists its config in ./data/whisper, and exposes Whisper on host port 10300 by default.

Prerequisites

• Docker with the Docker Compose plugin.

This repository does not include application source, package manifests, or language-specific build steps. Runtime behavior is defined by docker-compose.yml.

Run

docker compose up -d

Open http://localhost:3331.

Useful commands:

docker compose logs -f
docker compose pull
docker compose down

Configuration

Set overrides in your shell or in a local .env file before starting the stack. All variables below are optional.

Variable	Default	Purpose
PUID	1000	User ID for the Whisper container.
PGID	1000	Group ID for the Whisper container.
TZ	America/Vancouver	Time zone used by the Whisper container.
WHISPER_MODEL	auto	Whisper model loaded by faster-whisper.
WHISPER_LANG	auto	Language setting passed to faster-whisper.
WHISPER_BEAM	1	Beam size passed to faster-whisper.
WHISPER_PORT	10300	Host port forwarded to the Whisper service.

The agent talks to Whisper over the Compose network at whisper:10300. Changing WHISPER_PORT only changes the host-side port mapping.

No secret environment variables are required by the Compose file.

Contributing

Before changing runtime configuration, validate the Compose file:

docker compose config

If the change affects startup behavior, also run the stack locally and inspect logs:

docker compose up -d
docker compose logs -f