DBOS Transact

Open source
durable execution

Annotate your existing code in minutes, and make your applications fault-tolerant and observable by default. No orchestration server required.

Open source durable execution

"Makes it impossible to mess up."

Your existing code becomes fault-tolerant and observable by default. The simplest way to run code as it was meant to, no matter what.
scheduler.py
@DBOS.workflow()
def customer_workflow(scheduled_time: datetime,
customer_id: str):
   # Code to sync customer data
   ...

def on_customer_registration(customer_id: str):
   # Create a schedule dynamically per user
   # to run the customer_workflow
   DBOS.create_schedule(
       schedule_name=f"cus-{customer_id}-sync",
       workflow_fn=customer_workflow,
       schedule="0 * * * *",  # Runs every hour
       context=customer_id,
   )

Zero infrastructure

No new infrastructure

Simply connect to a Postgres-compatible database to store and retrieve state. No heavyweight orchestration server required.

Run anywhere

Run DBOS anywhere.

Portable and compatible with backend ecosystems, so you can run durable workflows wherever you want. On-premises or in the cloud.

Features

Like putting a PhD in your code

The open source DBOS Transact library simplifies workflow and AI agent idempotency, concurrency, observability and other backend complexities.

Reliable agents & workflows

Durable, idempotent workflows as code. Start, stop, resume, schedule, fork, and more.

Durable task queues

Scale easily with durable, concurrent task queues.

Agent & workflow observability

Emit OTel logs, traces, and metrics for easy troubleshooting.

4 billion workflows per day

Highly efficient reliability scales to handle large, complex use cases.

Run anywhere

Host your durable agents & applications anywhere.

Govern with DBOS Conductor

Your agent control plane - make workflow ops even easier.

Build with your favorite tools.
Deploy anywhere.

FAQs

Quick answers

Do I have to rewrite my application?

Not at all! DBOS is not a heavy framework that takes over your codebase. You only add DBOS decorators to the specific functions that actually need reliability.

How is this different from Temporal?

Temporal is an alternative durable execution solution to DBOS, but it requires you to deploy and maintain a separate, dedicated workflow orchestration server alongside your application.

DBOS Transact is a library that runs directly inside your existing application code and uses your existing Postgres database in which to store and recover workflow state and execution history.

You get enterprise-grade durability without an entirely new piece of infrastructure to manage.

Do I need cloud hosting to run DBOS Transact?

Nope! The DBOS Transact library is 100% open-source. You can run it on your local machine, or deploy it to your own cloud server.

Do I need DBOS Conductor to run DBOS Transact applications?

DBOS Conductor is not required to run durable workflows built with DBOS. But it does make them easier to manage.

DBOS Conductor is a workflow ops platform that gives you visibility into DBOS workflow and queueing state and history. It can alert you to failures, and allow you to stop, start, fork, pause workflows. It also automates high availability for your application instances, automatically resuming them if it detects outages.

The easiest way to build reliable backends

Build 10x faster

Just code your business logic – that’s it.
DBOS Transact makes it durable and observable for you.
Compatible with: Python Typescript Go Java Kotlin

Reliable by default

Add lightweight annotations to your app and make it resilient to any type of failure.

Observable by default

Generate logs and traces compatible with OpenTelemetry.

DBOS Transact

Open source
durable execution

The easiest way to make workflows and queues durable and observable.

Open source durable execution console