FAQ for Developers

Frequently asked questions about Pyck for developers

Audience: Programmer

How is data stored in Pyck?

Pyck uses a multi-tenant architecture with strict tenant isolation at the database level. Each tenant operates in a fully separated data space with no cross tenant data access. Within a tenant, multiple root level repositories (warehouses) can exist.

PostgreSQL serves as the primary data store. S3-compatible storage is used for files and attachments. The movement-based data model also enables time-travel queries, allowing you to query inventory state at any historical point in time.

Which database is used?

PostgreSQL is the primary database. S3-compatible storage is used as object storage for files and attachments.

What is the technology stack?

Component	Technology
API	GraphQL with Federation
Database	PostgreSQL
Event Streaming	NATS.io
Workflow Engine	Temporal.io
Authentication	OIDC/OAuth2 (e.g., Zitadel)
Containerization	Docker
Orchestration	Kubernetes

The logical foundation is built on these core concepts:

• Repositories — hierarchical containers for anything that can hold goods

• Movements — every inventory change as an executable instruction with full audit trail

• Movement Collections — ordered sequences of movements with forward-looking stock validation

• DataTypes — JSON Schema-based validation for flexible custom data

What are the Pyck framework building blocks?

The Pyck framework consists of three main components: the Core (backend services), the Web App, and the Mobile App.

Core: Backend services consisting of a GraphQL API (with Federation), event streaming (NATS.io), and workflow orchestration (Temporal.io).

Web App: Browser-based interface for managing warehouse operations, configuration, and monitoring.

Mobile App: Native mobile application for on-the-floor warehouse tasks such as picking, receiving, and inventory management.

Do data types need to be defined from scratch?

No. Pyck comes with predefined Entities (e.g., items, orders, customers, suppliers) that cover common warehouse concepts out of the box.

DataTypes are specializations of these Entities —— you define them yourself using an editor to add custom fields tailored to your business needs. Every Entity has a data field that is automatically validated against the custom JSON schema you can define, so your custom data is structured and type-safe without any backend code or database changes.

How are processes defined and modeled?

Pyck uses Temporal.io as its workflow engine. Logistics processes are defined as Temporal workflows — primarily in Go, as this is the Pyck SDK that is currently available, though Temporal itself supports multiple languages. The Pyck Go SDK is optional and serves as a convenience layer; workflows can also be implemented without it.

For Pyck-specific workflow concepts, see:

• Workflows (Concepts)

• Workflows (Development)

For details on authoring Temporal workflows, refer to the Temporal Go SDK documentation.

What programming language is the code written in?

Go. All backend services are written in Go.

Maintainability is supported by:

• API-first design — the GraphQL schema drives implementation

• Microservice architecture — clear separation of concerns

How do concurrent processes affect each other?

Pyck is designed to handle concurrent API requests without conflicts. Database transactions and stock validations ensure operational consistency, even under concurrent load.