This blog post is a public copy of our monthly newsletter. If you’d like to receive upcoming updates by email, you can subscribe to our newsletter here.
Highlights from January include new integrations, hands-on tutorials, and builder spotlights from folks pushing Temporal in creative directions. You’ll also find answers to real questions from the field and a few chances to catch us live.
For more information and other things we’ve been working on, just keep reading! And as always, we’d love to hear from you. Feel free to share feedback in our Community Slack.
Recently shipped#
Now in Public Preview, these new integrations help teams build production-ready AI apps and agents with added durability and observability.
- AI SDK by Vercel: The AI SDK by Vercel helps TypeScript developers ship AI features quickly. With the new Temporal integration, you can add durable state, retries, and long-running workflows to AI agents and apps without changing how you write your code.
- Braintrust: The Braintrust and Temporal integration provides end-to-end visibility into AI workflows, including model calls, decisions, retries, tool usage, and prompts, so teams can debug faster and improve AI behavior.
- AI Cookbook recipes: New recipes give you production-ready patterns for building durable, observable AI applications and features.
- API and UI for viewing Audit Logs in Temporal Cloud is now generally available! This will help you validate changes in your environment and easily see who made changes to what.
- Temporal Docker images are being restructured with the upcoming release of 1.30.0. These changes improve security and slim down the existing collection. Users can still leverage old versions of the Docker images if needed.
Builder spotlight#
We’ve got three cool projects to showcase this month. First, check out Juan José Orjuela Castillo’s post about deploying Temporal using AWS ECS — written in Spanish! Check out his post and see how Temporal can work in tandem with Terraform and Datadog.
Next, the team over at Galaxy loves Temporal but hates leaving the terminal. So they created tempo, a TUI that can be used to browse Workflows, inspect Histories, switch Namespaces, and a ton more. Check out the repo and try it for yourself!
Sophia Barnes recently described the multi-agent system she designed for a reinsurance company. This system uses specialized agents to extract catastrophe loss data from non-standardized Excel submission packs. Each agent runs as a Temporal Workflow, using Activities to handle LLM calls, while another Workflow coordinates agent interactions.
How to Temporal#
We’ve just released two brand new tutorials as part of our growing library of AI integration resources. The tutorials show how Temporal adds durability and reliability to MCP servers, even when things go wrong:
Questions from the field#
John Koehn mentioned in #support-community that one of their Workflows was suddenly reporting a large number of Workflow Tasks being closed. Tamara Henson explained that this can occur when a Workflow schedules too many Activities or does too much work in a single Task. Tihomir Surdilovic elaborated on this, mentioning relevant metrics and advice on how to resolve it.
Nikita Pozniakov asked about best practices for horizontally scaling Workflow processing. Maxim Fateev explained that there’s no specific limit on Workflows per Worker — the real constraint is on the number of Workflow Tasks processed per second. Stephen Montague advised scaling horizontally first for resilience, then vertically for utilization, providing helpful links to Temporal’s deployment documentation.
Join us live#
Connect with our team at upcoming events to see demos from real humans. We’d love to show you exactly how we’re powering mission-critical systems at world-class organizations. That includes Replay ‘26, our annual conference, where you can hear directly from OpenAI, Replit, Cursor, NVIDIA, Netflix, and more. Use code SPOTLIGHT75 for 75% off.