CopyCat

CopyCat is a Discord bot that lets you send messages and voice clips as yourself — using your own name and avatar — via Discord's webhook system.

Commands

Command	Description
/s	Send a message (with optional attachment) as yourself through a webhook. Supports replying to other messages.
/tts	Generate a text-to-speech voice clip and post it as yourself. Choose from dozens of neural voices (Piper TTS) or novelty voices like Morshu and CABAL.

TTS voices

Voices are split into two backends:

• Piper TTS — neural, offline voices covering many languages. Models are downloaded to a mounted volume at container startup (/app/piper). Only voices whose .onnx model file is present on disk appear in autocomplete.
• Python bridge — Morshu (MorshuTalk) and CABAL (TiberianSunCABAL-Talk). Require their vendor repos to be cloned into Tools/Tts/vendors/ inside the container. Only appear in autocomplete when their vendor directory is found.

Self-hosting

Requirements

• Docker and Docker Compose.
• A Discord application with a bot token and client ID.

Setup

1. Create a .env file in the project root:

CLIENT_ID=your_client_id_here
BOT_TOKEN=your_bot_token_here

To enable pk;admin commands (raising member limits, etc.), also set:

ADMIN_ROLE=your_admin_role_id_here

2. Build and start all services:

docker compose build
docker compose up -d

3. Register slash commands (run once, or after command changes):

docker compose run --rm register-commands

4. View logs:

docker compose logs -f bot

Piper voice models

Set COPYCAT_TTS_DOWNLOAD_VOICES=1 on the bot service (it is set by default in docker-compose.yml) to automatically download Piper voice models into /data/piper on the host at container startup. Only voices with a downloaded model file will appear in the /tts autocomplete list.

To add a custom voice model, place the .onnx and .onnx.json files in /data/piper on the host (mounted into the container at /app/piper).

CABAL audio assets

The CABAL voice requires copyright-encumbered game audio files (.aud) that are not included in the repository. Supply them by placing the files in /data/cabal on the host before starting the bot:

/data/cabal/
  01-n400.aud
  nod-02.aud
  nod-07b.aud
  ... (other .aud files from the cabal/ directory)

The directory is bind-mounted into the container at /app/cabal. Without these files the CABAL voice will fail at runtime; all other voices are unaffected.

Data storage

• PostgreSQL — all persistent data, stored in /data/db on the host.
• Redis — internal state and transient data.
• Piper voices — model files in /data/piper on the host (mounted into the container at /app/piper).
• CABAL audio — .aud game audio files in /data/cabal on the host (mounted into the container at /app/cabal). Must be supplied manually (see above).

Architecture

Service	Language	Role
bot	C# (.NET)	Discord bot: handles slash commands, proxying, TTS
gateway	Rust	Connects to Discord gateway and forwards events to the bot
migrate	Rust	Runs database migrations on startup
register-commands	Rust	Registers slash commands with Discord
db	PostgreSQL 17	Persistent data store
redis	Redis	Internal state and caching

Development

See dev-docs/ for build instructions, local development setup, and service-specific configuration.

License

With the exception of the Myriad library, this project is licensed under the GNU Affero General Public License, Version 3. See the COPYING file for the full license text.

Licensing information for the Myriad library can be found in Myriad/COPYING.