This morning at Replay 2026 in San Francisco, we announced exciting new capabilities in Temporal to enable developers to build durable, reliable applications and agents.
Read on to learn what's new, and reach out to us in Slack to share your questions and thoughts. We'd love to hear from you.
Day 1 Announcements#
Develop and deploy faster with new capabilities & primitives#
We've released new capabilities to accelerate your development with Temporal.
Deploy Workflows more quickly and improve costs with Serverless Workers#
We're excited to share that you can now run Temporal Workers on serverless compute, starting with AWS Lambda. Serverless Workers reduce time to get started by removing infrastructure planning overhead and eliminating the need for autoscaling strategy design. Temporal Cloud now automatically invokes, scales, and gracefully shuts down (scaling to zero, if appropriate) Temporal Workers on your behalf based on workload volume and metrics. Serverless Workers are now in Pre-release.
Write lightweight, durable jobs with Standalone Activities#
Today, we introduced a new primitive in Temporal: the Standalone Activity, which enables Activities to run on their own, not just as steps inside a Workflow. Standalone Activities let you improve the durability and debuggability of job processing, and eliminate complex queuing and retry logic by adopting the same model you already know and love. When your use case outgrows a single step, you can leverage that same Activity inside a Workflow, no rewrite required.
Standalone Activities are now available in Public Preview for the Go, Python, and .NET SDKs, and in Pre-release for the Java and TypeScript SDKs, and supported in Temporal Cloud.
Build better AI experiences, with more control#
Our recent updates for developers building AI agents and applications let them operate more efficiently and create richer end-user experiences.
See real-time LLM responses and tool call updates#
We've released Workflow Streams: durable streaming that uses Temporal's Signal & Update primitives. With Workflow Streams, you get token batches and application-level updates to power responsive UIs, live monitoring, and guardrails. Workflow Streams is designed for real-time user output while preserving Temporal's reliability model. Now in Public Preview.
Reliably handle large AI Inputs and Outputs with External Payload Storage#
With Temporal's new External Storage feature, you can more efficiently and reliably store and retrieve large payloads from Workflows and Activities. It's easy to wire up Amazon S3 support to your Workflows, or you can provide a custom storage Driver. External Storage is especially useful for agentic and AI data processing applications. It's now in Public Preview for Python and Go.
Integrate with a rich and growing ecosystem of AI frameworks#
We continue to add AI partners and integrations to the Temporal ecosystem so you can build AI agents and apps with the tools you love.
Introducing the Temporal AI Partner Ecosystem#
Our goal is to meet developers where they are building AI agents and applications. That's why we're investing in making it easier for partners to build Temporal integrations that help developers deploy AI applications and agents faster and more reliably. You can learn more and submit your interest to build an integration here.
Build agents faster with the new Google ADK integration#
Temporal's Google ADK integration makes agents resilient, scalable, and production-ready. LLM calls and tool executions run as Temporal Activities, giving each step automatic retries and recovery out of the box. If anything fails, your agent picks up exactly where it left off.
Develop durable sandbox agents with OpenAI Agents SDK sandbox support#
Temporal's Python SDK now integrates with the Agent SDK's new sandbox functionality, which lets agents execute in isolated environments. Temporal gives your sandbox agents durability, easy session management, and more. Also, the general integration with the OpenAI Agents SDK is now Generally Available.
AWS AI Competency#
Temporal has earned the AWS AI Competency in the Agentic AI category, recognizing our proven expertise in helping organizations build and run reliable, production-grade AI agents on AWS. It's a distinction held by a select group of AWS Partners with verified technical depth and real customer outcomes to back it up.
What's next#
If you have any questions about these new releases, let us know in the Temporal Community Slack. And stay tuned – tomorrow, we’ll be back with more product updates from Replay 2026!