Development workflow¶
Tinybird follows a branch-based development workflow. You edit project files in your local repo that is connected to Git, test changes in Tinybird Local or a Cloud Branch, and deploy a new version to your Cloud Workspace.
Use this workflow when you change Data Source schemas, iterate on Pipe logic, add Endpoints, or validate connector ingestion before it reaches production.
Choose an environment¶
| Environment | Use it for | Example |
|---|---|---|
| Tinybird Local | Fast local iteration and CI checks with fixture data. | Change a Pipe query, run tb build, and call the local Endpoint until it returns the expected result. |
| Cloud Branches | Testing with production-like data without changing production. | Check out a Git branch, run tb build, change a Data Source schema, and query the branch before deploying. |
| Staging deployments | Validating the deployment that can become live. | Run tb --cloud deployment create --wait, write compatible test events to staging, then promote. |
Local and branch environments are for development. Deployments are for moving a validated project version toward production.
Develop in a branch¶
Choose the development loop that fits the data you need:
- Local development: Run Tinybird Local with fixture data and local APIs.
- Cloud Branches: Use a Cloud environment for production-like data, connector validation, preview apps, and pull request review.
Test your changes¶
Test Tinybird changes as part of your application workflow, the same way you test the rest of your app. Test locally with fixtures, or test against a Cloud Branch when you need production-like data. You can also validate your data project independently with fixture files, data quality tests, and deployment checks. See Test your project.
Deploy to your Cloud Workspace¶
Push your changes to Tinybird Cloud with automatic schema migrations. Run a deployment check first when the change affects Data Source schemas, Materialized Views, engine settings, or connector configuration.
- Examples: See concrete workflows for Data Source evolution, Pipe iteration, SDK projects, and deployments.
- Evolve Data Sources: Understand how schema, engine, and Materialized View changes are migrated during deployments.
- Manual deployment: Deploy directly from the CLI.
- CI/CD: Automate testing and deployment with GitHub Actions or GitLab CI.
Examples¶
See Development workflow examples for common loops with datafiles, the TypeScript SDK, and the Python SDK.
Core concepts¶
- Tinybird Local explains the local runtime, services, installation, Docker usage, persistence, and local Tokens.
- Branches explains the Branch resource model and data isolation semantics.