tycostream

tycostream turns Materialize views into GraphQL APIs, so you can quickly build real-time, type-safe apps. Like Hasura, but for streaming use cases.

Key features:

• Subscriptions — Subscribe to any view and get updates over WebSockets
• Triggers — Fire webhooks when data meets specific conditions

Learn more about why I built this →

---

Quickstart

1. Configure connection details

# Create .env file with your database connection
echo "DATABASE_HOST=localhost
DATABASE_PORT=6875
DATABASE_USER=materialize
DATABASE_PASSWORD=materialize
DATABASE_NAME=materialize
GRAPHQL_PORT=4000
GRAPHQL_UI=true
SCHEMA_PATH=./schema.yaml
LOG_LEVEL=debug" > .env

# If Materialize runs on your host machine and you're using Docker:
# macOS/Windows: Change DATABASE_HOST to host.docker.internal
# Linux: Change DATABASE_HOST to 172.17.0.1

2. Create a schema file

You need to tell tycostream which tables and views to expose in your API. For example, given a trades table like this:

CREATE TABLE trades (
  id INT,
  instrument_id INT,
  side TEXT,           -- 'buy' or 'sell'
  quantity INT,
  price NUMERIC,
  executed_at TIMESTAMP
);

You would need a schema.yaml file like this:

enums:
  side:
    - buy
    - sell

sources:
  trades:
    primary_key: id
    columns:
      id: Integer
      instrument_id: Integer
      side: side              # mapped to enum
      quantity: Integer
      price: Float
      executed_at: Timestamp

You can use the bundled generator script to create schema files directly from your Materialize instance. For the trades table example above, the command would look like this:

curl -sL https://raw.githubusercontent.com/tycoworks/tycostream/main/scripts/generate-schema.sh | \
  bash -s -- -e side "buy,sell" -s trades -p id -c side:side > schema.yaml

3. Start tycostream

docker run -p 4000:4000 --env-file .env \
  -v $(pwd)/schema.yaml:/app/schema.yaml \
  ghcr.io/tycoworks/tycostream:v0.1.0-preview

4. Test your API

Visit http://localhost:4000/graphql to explore your GraphQL API with the built-in UI (enabled by GRAPHQL_UI=true in .env).

---

Building from Source

Getting Started

For development, customization, or running the demo:

git clone https://github.com/tycoworks/tycostream.git
cd tycostream
npm install
npm run build
npm start

Running the Demo

Once you've cloned and installed:

npm run demo

This starts a complete example with:

• Live market data streaming
• Real-time position tracking
• AG-Grid frontend with WebSocket subscriptions

Visit http://localhost:5173 to see the demo UI.

tycostream.demolow.mov

---

What's Next

See the roadmap for planned features.