Development Workflows

Tinybird supports three development workflows. Choose based on your team size, infrastructure, and iteration speed needs.

Workflow Comparison

Workflow	Best for	Requires	Data
Local (`dev_mode=local`)	Solo dev, fast iteration, offline work	Docker	Fixtures or manually appended
Branch (`dev_mode=branch`)	Team collaboration, production-like testing	Cloud workspace	Optional copy from production
Cloud direct	Simple projects, quick prototyping	Cloud workspace	Production data

Recommended: Branch Workflow

For most projects, use dev_mode=branch. It provides isolated environments backed by Tinybird Cloud, with optional access to production data.

json

{
  "dev_mode": "branch"
}

Create a git branch for your feature
Run tb dev — a Cloud branch is created automatically from the git branch name, file changes are watched and auto-rebuilt
Develop and test: tb endpoint data <pipe_name>
Push, create PR — CI runs tb --cloud deploy --check
Merge — CD runs tb --cloud deploy

See rules/branch-development.md for details on branch tokens and --last-partition.

Local Workflow

Use dev_mode=local for fast iteration without network dependencies. Good for developing SQL logic and testing with fixture data.

json

{
  "dev_mode": "local"
}

Start Tinybird Local: tb local start
Run tb dev in a new terminal — watches files and auto-rebuilds
Append test data: tb datasource append <name> --file fixtures/<name>.ndjson
Test endpoints: tb endpoint data <pipe_name>
Deploy when ready: tb --cloud deploy

See rules/local-development.md for Tinybird Local commands and troubleshooting.

Cloud Direct Workflow

For simple projects or quick prototyping, you can work directly against Cloud. Use tb --cloud deploy to deploy, or the two-step process for explicit confirmation:

tb --cloud deployment create --wait
tb --cloud deployment promote

Or the combined shorthand:

tb --cloud deploy

Choosing a Workflow

Starting a new project? Start with Local for fast bootstrapping, switch to Branch when you need production data or team collaboration.
Team project with shared workspace? Use Branch. Each developer gets an isolated environment.
Quick prototype or demo? Cloud direct is fine.
CI/CD pipeline? Use Tinybird Local for CI build/test, then tb --cloud deploy for production. See rules/ci-cd.md.

Testing Endpoints

Use tb endpoint data to test endpoint output:

tb endpoint data my_endpoint
tb endpoint data my_endpoint --start_date 2024-01-01 --end_date 2024-01-31

Use tb endpoint data, not tb pipe data. The endpoint data command calls the endpoint as an API consumer would, including parameter validation and output formatting.