Zigflow provides a simple and declarative way to define and manage Temporal workflows using the CNCF Serverless Workflow specification. It enables low-code and no-code workflow creation that's easy to visualise, share, and maintain, without sacrificing the power and reliability of Temporal.
- ✅ CNCF Standard – fully aligned with Serverless Workflow v1.0+
- ✅ Low-code & Visual-ready – ideal for UI workflow builders and orchestration tools
- ✅ Powered by Temporal – battle-tested reliability, retries, and state management
- ✅ Kubernetes-native – includes a Helm chart for easy deployment
- ✅ Open & Extensible – customize, extend, and contribute easily
Ziggy is a tardigrade and Temporal's official mascot. They're a microscopic animal that is basically indestructible.
Sound familiar?
Define a workflow declaratively in YAML:
document:
dsl: 1.0.0
namespace: MoneyTransfer # Mapped to the task queue
name: AccountTransferWorkflow # Workflow name
version: 0.0.1
title: Money Transfer Demo
summary: Temporal's world-famous Money Transfer Demo, in DSL form
do:
- queryState:
listen:
to:
one:
with:
# ID maps to the query name in Temporal
id: transferStatus
# Temporal query - used to make read requests
type: query
# The data returned from the query - for application/json, this must be a string so Go interpolation works correctly
data:
approvalTime: ${ .data.stateApprovalTime }
chargeResult:
chargeId: ${ .data.stateChargeId }
progressPercentage: ${ .data.stateProgressPercentage }
transferState: ${ .data.stateTransferState }
workflowStatus: ${ .data.stateWorkflowStatus }
- setup:
set:
idempotencyKey: ${ uuid }
stateApprovalTime: 30
stateChargeId: ${ uuid }
stateProgressPercentage: 0
stateTransferState: starting
stateWorkflowStatus: ""
- validate:
call: http
with:
method: post
endpoint: http://server:3000/validate
- updateState:
set:
stateProgressPercentage: 25
stateTransferState: running
- withdraw:
call: http
with:
method: post
endpoint: http://server:3000/withdraw
headers:
content-type: application/json
body:
amount: ${ .input.amount }
attempt: ${ .data.activity.attempt }
idempotencyKey: ${ .data.idempotencyKey }
name: ${ .data.workflow.workflow_type_name }
- updateState:
set:
stateProgressPercentage: 50
- deposit:
call: http
with:
method: post
endpoint: http://server:3000/deposit
headers:
content-type: application/json
body:
amount: ${ .input.amount }
attempt: ${ .data.activity.attempt }
idempotencyKey: ${ .data.idempotencyKey }
name: ${ .data.workflow.workflow_type_name }
- updateState:
set:
stateProgressPercentage: 75
- sendNotification:
call: http
with:
method: post
endpoint: http://server:3000/notify
headers:
content-type: application/json
body:
amount: ${ .input.amount }
fromAccount: ${ .input.fromAccount }
toAccount: ${ .input.toAccount }
- updateState:
set:
stateProgressPercentage: 100
stateTransferState: finishedRun it through Zigflow:
zigflow -f ./path/to/workflow.yamlThis builds your Temporal workflow and runs the workers — no additional Go boilerplate required.
You can now run it with any Temporal SDK.
- Task Queue:
zigflow - Workflow Type:
example
Telemetry helps us understand if people are actually using Zigflow. Stars are great, but a tiny bit of usage data gives us a clearer picture of what’s happening out there.
Zigflow creates a UUID (stored in ~/.config/zigflow) and sends just two things:
- that anonymous ID
- the Zigflow version you’re running
That's it.
Telemetry is optional and easy to turn off. You can disable it by setting the environment variable:
DISABLE_TELEMETRY=trueor by running Zigflow with:
--disable-telemetryContributions are welcome!
All commits must be done in the Conventional Commit format.
<type>[optional scope]: <description>
[optional body]
[optional footer(s)]
Made with contrib.rocks.
Distributed under the Apache-2.0 license
© 2025 Zigflow authors
