Heather Anne Ritchie

@DBOS_Inc project seems to be really cool, you can easily write background agentic workflows in Python with checkpoints stored in PostgreSQL, durable execution, no external message queues, integrates with @pydantic AI.

Do you like databases? Do you want to hear two database professors rant about them? Do you need one of those professors to have a Turing Award for databases? If yes, then join Mike Stonebraker and I next Wed Dec 10 @ 1:00pm EST for database hot takes: dbos.dev/webcast-2025-i…

After connecting your @DBOS_Inc application to Supabase, you can use the Supabase console to view your DBOS system tables

We usually think of workflows as centralized–there’s a big central orchestrator telling different workers what to do. But what if they don’t have to be? What if your app servers could run whatever workflows they need without central orchestration? That would be a path towards radically simpler infrastructure. But building decentralized workflows is inherently challenging. Servers need to cooperate to efficiently schedule and execute workflows, but they can’t directly communicate with each other. How do we make this work? Check out my new blog post! 👇

That moment you realize the fire wasn’t in staging.

@qianl_cs @jedberg ...fwiw I gave him a hard time for not having an alternative... 😆

The outage is still going on... Our CEO @jedberg is sitting in the dark today because his smart bulbs depend on us-east-1.

Cool new feature–you can now configure DBOS Python with a custom serializer! By default, DBOS uses pickle to serialize workflow state to store in Postgres. But for better interoperability or easier integrations, many of you want to use different serializers, like JSON, Pydantic, or anything else in the ecosystem. Now, you can write any serializer you need, plug it in, and it just works.

My colleague @DevHawk is presenting today at the Seattle Postgres User Group (SEA PUG)! If you're in the area, drop by and say hi :) "Durable Computing 101 with Postgres and Python" meetup.com/seattle-postgr…

We had our first job posting from a customer on our Discord! And yes, I immediately told them to cross-post in @eatonphil Software Internals channel too. Awesome to see teams hiring specifically to expand their DBOS use cases. Link to our Discord below.

Pydantic AI + DBOS durable agent integration is now live! You can now make Pydantic agents durable by wrapping them with DBOSAgent. I've been really impressed by Pydantic AI. It's flexible and dynamic without requiring a predefined DAG. Execution is automatically traced and can be exported to Logfire for observability. The public API is straightforward to extend with new functionality. Most importantly, the @pydantic team (especially @DouweM @david_montague_ @samuelcolvin) is very responsive and supportive of contributors. We designed the integration to be both robust and flexible: - You can use either sync or async Python. - The agent's main loop is automatically a DBOS workflow. - External model calls and MCP communication are DBOS steps with configurable retries. - You can decide how to handle custom tools: either as a step, a normal Python function (avoiding checkpoint overhead), or a function that enqueues background tasks.

We're currently on the Hacker News front page! Grateful to @jedberg for sharing his experience running queue services at Reddit, and for reminding us why durable queues matter. Spoiler alert: "If the queue itself went down... we could just lose votes, or comments, or submissions. Did you ever think ‘I know I voted on that but it’s gone!’ on Reddit? That’s why."

This isn’t another AI webinar. It’s the blueprint for workflows that don’t quit. Join DBOS & Bauplan to learn how to build autonomous and resilient AI workflows. oglink.xyz/aq1Jlxbbvj

DBOS ensures every step in your workflow runs exactly once, even in the face of crashes or network issues. Thanks to time-travel logs, you can restore your application/database to any previous point (before a bug, a bad deploy, or an attack) and replay forward, minimizing impact.

Congratulations @yutori_ai on the launch! As a startup founder, few moments compare to watching your customers succeed and knowing your product played a part in their journey. Yutori has been an amazing design partner, using DBOS to build the workflows powering scouts. They’re a smart team with a compelling vision for the future of agents. Now that they're live, I can't wait to see how users experience what they've built!

@petereliaskraft @yutori_ai 👏👏👏

@deviparikh Congratulations Devi! Can’t wait to try it.

It's here! Introducing Scouts by Yutori. Scouts is like having a team of agents monitoring the web for information that matters to you. We're letting more users in everyday. Join the waitlist!

@abhshkdz Congratulations on a big day! (And great video, team.)

We're excited to launch Scouts — always-on AI agents that monitor the web for anything you care about.

3/ Compared to AWS Step Functions or Temporal, DBOS stands out with easy integration and no ongoing hosting fees. Durable execution is finally accessible for every developer, with less friction and more flexibility. 🔗Learn more: dbos.dev/blog/durable-e…

🎙️ How does durable execution make AI more reliable? Ryan Donovan is joined on the podcast by @jedberg, CEO of @DBOS_Inc, and @qianl_cs, co-founder, to dive into its use cases, and how it's changing the future of debugging and observability for tech. stackoverflow.blog/2025/05/20/dur…

One cool thing about implementing workflows in Postgres is that you can access them from anywhere. All the workflow’s state is in a Postgres table, so you can start new workflows or interact with running workflows with SQL. So today we released DBOS Clients, which let you programmatically interact with your workflows from anywhere, through Postgres. You can use clients to enqueue new workflows, send messages to your running workflows, or check your workflows’ status. All you need to instantiate a client is a Postgres connection, so you can add them to any service. We’re going to keep expanding what clients can do over the next few weeks, including more tooling for workflow introspection and management, so stay tuned!