Patrizio

A Delta Chat bot for group chats, built with Go. Patrizio responds to messages based on configured keyword filters, inspired by Miss Rose on Telegram. Its main aim is to be used in friend groups or with people known. Differently from Miss Rose, it does not serve as a bot moderator, since Delta Chat groups have no admin/member structure.

Why Patrizio?

While it is not properly the penguin in the image, it comes from a funny Internet meme which was trending some time ago. Since I am always out of ideas for good names, I thought it was a good pick. After all, who does not like penguins?

Quick Start

1. Build

make build

2. Initialize the bot

Configure the bot with a Delta Chat email account:

./patrizio init bot@example.com YOUR_PASSWORD

3. Get the invite link

Share this link so users can contact the bot:

./patrizio link

4. Run the bot

./patrizio serve

The bot will connect to Delta Chat and start processing messages. Add it to a group to use keyword filters, or send it a direct message to get help text.

Configuration

Configuration is done via environment variables prefixed with PATRIZIO_:

Variable	Default	Description
PATRIZIO_DB_PATH	./patrizio.db	Path to the SQLite database file
PATRIZIO_MEDIA_PATH	./media	Directory where media files are stored
PATRIZIO_LOG_LEVEL	info	Log level

The bot's Delta Chat account data is stored in a platform-specific config directory (e.g. ~/.config/patrizio/ on Linux), overridable with --folder:

./patrizio --folder /custom/path serve # --folder is optional, config will be saved in your user home otherwise

Docker

Build and run with Docker (this assumes you've already initialized your bot instance):

make docker-build
docker run -v patrizio-data:/data patrizio -f /data serve

Docker compose

If you prefer an orchestrated solution, you can use the included Docker Compose, which will build the image from the source and serve it as it is. This is particularly useful if you do not want to rely on the provided Docker images, or if you want to apply local edits. Before proceeding folder, a ./data folder is mandatory in order for this setup to work properly. It is also necessary to give the right permissions to the folder, so that the non-root user in the container has access to it:

mkdir -p ./data/{media,db}
sudo chown -R 65532:65532 ./data # 65532 is the UID of the image.

Once setup, you have to populate the folder properly, which you can do with:

make docker-build
docker run --rm -v ./data:/data patrizio -f /data init
docker run --rm -v ./data:/data patrizio -f /data link

Then you can run it with:

docker compose up --build -d

Note

The included Docker Compose has a built-in log rotation (up to 10MB)

Development

Setup

A Make directive is available for a quick project setup. Note that this will not install any dependencies, i.e. Golang, Golint-ci and Pre-commit.

make project-setup

Makefile Targets

Target	Description
make project-setup	Setup project related hooks (doesn't install new software)
make build	Compile the binary
make run	Run the bot in serve mode
make test	Run all tests
make lint	Run golangci-lint
make docker-build	Build the Docker image
make migrate	Run pending database migrations
make migrate-create NAME=<name>	Create a new migration file
make sqlc	Regenerate Go code from SQL query files
make doc-setup	Command to only setup documentation (included in project setup)
make doc-build	Builds the documentation and output in site directory
make doc-local	Serves the documentation locally, at localhost:8000
make clean	Remove build artifacts

Database Migrations

Migrations can be created with:

make migrate-create NAME=add_filters_table

This creates a new .sql file in migrations/. Edit it, then run:

make migrate

Migrations are also run automatically on bot startup.

SQL Queries

SQL queries live in queries/ as .sql files. After editing, regenerate the Go code:

make sqlc

Generated code is written to internal/database/queries/.

AI disclamer

As can be see by the openspec folder, the heavy lifting of this project has been done by using AI (Claude CLI, in particular since I try to stick to local models hosted via Ollama as much as possible). This means, bootstrapping the project and adding the very first feature. I would never had enough time to learn all the Delta Chat RPC basics and to start the project. I understand someone might not be OK with it, but by using it and contributing, you accept this fact. Other AI assistance tools will be used during the development, in particular with the aim to explore and learn Delta Chat RPC and AI tools.

More?

You've made 'till here and want to read more? Head over the project webiste for more!

License

AGPL-3.0-or-later. See LICENSE for the full text.