Don't make Claude do the same work twice
Claude Agent SDK runs the agent loop. Kitaru adds the durable runtime around a completed invocation — checkpointed results, artifacts, replay boundaries, and waits.
Jun 1, 20268 mins
23 posts with this tag
Claude Agent SDK runs the agent loop. Kitaru adds the durable runtime around a completed invocation — checkpointed results, artifacts, replay boundaries, and waits.
LangGraph keeps graph state, threads, and interrupts. Kitaru adds the durable workflow around the graph call — replay boundaries, durable waits, and inspectable runs.
The OpenAI Agents SDK stays the harness; Kitaru adds the runtime around it — durable workflow waits, replay boundaries, and inspectable execution history.
Armin Ronacher's Absurd and Kitaru arrived at the same answers on replay semantics, ephemeral compute, and an agent-legible runtime. Here's why that matters.
Meet Kitaru — open source durable execution for Python agents, built by the ZenML team. Crash recovery, human-in-the-loop, and replay from any checkpoint.
Kitaru is live: open-source infrastructure platform for running Python agents in production.
We spent five years building ML pipeline infrastructure. Then agents showed up and we realized the next problem needed a new tool — not an extension of the old one.
Tracing shows you what went wrong. But what if you could go back, fix the input, and resume from where it failed — without re-running everything?
I rebuilt zenml.io — 2,224 pages, 20 CMS collections — from Webflow to Astro in a week using Claude Code and a multi-model AI workflow. Here's how.
AI agents fail — they timeout, hit rate limits, crash on bad API responses. Without durable execution, every failure means starting over from scratch.
In this Langfuse vs Phoenix guide, we conclude which open-source framework fits your LLMs stack by comparing features, integration, and pricing.
Two open-source contributors describe how they built a new onboarding experience for the ZenML Quickstart native to VS Code.
Master cloud-based LLM finetuning: Set up infrastructure, run pipelines, and manage experiments with ZenML's Model Control Plane for Microsoft's latest Phi model.
Master cloud-based LLM finetuning: Set up infrastructure, run pipelines, and manage experiments with ZenML's Model Control Plane for Meta's latest Llama model.
We've open-sourced our new dashboard to unify the experience for OSS and cloud users, although some features are initially CLI-only. This launch enhances onboarding and simplifies maintenance. Cloud users will see no change, while OSS users can enjoy a new interface and DAG visualizer. We encourage community contributions to help us expand and refine this dashboard further, looking forward to integrating more features soon.
Community member Marwan Zaarab explains how and why he built a VS Code Extension for ZenML.
ZenML secures an additional $3.7M in funding led by Point Nine, bringing its total Seed Round to $6.4M, to further its mission of simplifying MLOps. The startup is set to launch ZenML Cloud, a managed service with advanced features, while continuing to expand its open-source framework.
This week I spoke with Emeli Dral, co-founder and CTO of Evidently, an open-source tool tackling the problem of monitoring of models and data for machine learning. We discussed the challenges around building a tool that is both straightforward to use while also customizable and powerful.
We put together a list of 48 open-source annotation and labeling tools to support different kinds of machine-learning projects.
As we outgrew our initial template Github Action workflow, here's the five things we added to our Github Action arsenal to fit our growing needs: Caching, Reusable Workflows, Composite Actions, Comment Triggers and Concurrency Management.
We recently reworked a number of parts of our CLI interface. Here are some quick wins we implemented along the way that can help you improve how users interact with your CLI via the popular open-source library, rich.
An overview of some of the capabilities that ZenML will unlock for our users.
