July 2026 Summaries

Developers often face challenges with traditional infrastructure management, such as provisioning, scaling, and managing upgrades, which can be resource-intensive and costly. To address these issues, Temporal Serverless Workers have been introduced for Google Cloud Run, simplifying the process of running and scaling workers without the need for extensive infrastructure planning or autoscaling strategies. Google Cloud Run, a fully managed platform, now supports Serverless Workers alongside its Worker Pools, which are tailored for long-running tasks. Temporal's integration with Cloud Run enhances the operational model by offering a responsive and efficient autoscaling algorithm that adjusts capacity dynamically based on workload metrics, eliminating the need for manual setup. This setup allows developers to deploy worker code with ease, and Temporal manages all scaling activities, including scaling to zero when necessary. Serverless Workers for Google Cloud Run is currently available as a pre-release product, inviting interested users to explore its capabilities.

Temporal’s July update highlights new capabilities for building and operating durable applications, including pre-release Serverless Workers for Google Cloud Run, which automatically invokes, scales, and shuts down Workers based on workload demand. Public-preview integrations with LangGraph and LangSmith allow developers to retain those frameworks while using Temporal for production execution, while generally available billing APIs, billable-action metrics, and Datadog and Vantage integrations improve Temporal Cloud cost visibility. A Workers UI update adds monitoring details such as heartbeats, deployments, CPU usage, cache-hit ratios, and task capacity. The community spotlight features Jeff Romine’s full-stack e-commerce demo, which uses Temporal workflows for every business-state transition and combines TypeScript, Next.js, and Cassandra. New Validated Patterns and technical resources cover Temporal Nexus for .NET and migration from ARQ for agentic workflows, while community discussions address Python request-ID propagation through interceptors and memory-growth mitigation for Temporal History pods. The update also promotes August webinars, livestreams, and an in-person appearance at PyCon in Brisbane.

Houman Kargaran, an Engineering Lead at ANZ Bank, discusses the complexities of processing Pub/Sub messages before the advent of Temporal's Standalone Activities, highlighting the challenges faced by engineers in managing retries, state tracking, and visibility when processing messages from a queue. Standalone Activities, now in Public Preview, offer a streamlined approach by allowing a single activity to run independently, acting as a durable bridge between Pub/Sub and Temporal, thus simplifying event handling by ensuring that each file ingestion remains visible and traceable. This method reduces overhead by eliminating the need for a full Workflow and is particularly useful for executing single functions with built-in retries and timeouts, such as sending emails or processing webhooks, while keeping the event handler decoupled from the processing logic. The article emphasizes the benefits of Standalone Activities in maintaining a thin event handler and suggests that further exploration into the first Activity within the Workflow will be covered in a subsequent article.

Houman Kargaran, Engineering Lead at ANZ Bank, discusses building a PII-compliant complaint ingestion pipeline using Temporal as the orchestration layer, ensuring that sensitive data does not leave the enterprise's infrastructure. The pipeline processes CSV files containing customer complaints by detecting PII, classifying them with an internal ML model, redacting sensitive information, and enabling semantic search. Key requirements include maintaining data within the enterprise, ensuring no complaints are lost, and making the system resilient, consistent, and observable. Temporal is chosen for its ability to provide durable execution, retries, and a detailed audit trail, aligning with compliance needs. The article outlines how Temporal's capabilities address both functional and non-functional requirements, including resilience, consistency, availability, and observability, and describes how each step in the pipeline is managed as a discrete Temporal Activity with specific retry policies. The series will continue to explore each layer of the system in detail, explaining the integration of Temporal with Pub/Sub, security boundaries, and activity chaining.

Zscaler architects Kartick Krishnachetty Ravi and Abhishek Mitra describe building an automation platform that uses Temporal Workflows as a durable orchestration layer above a heterogeneous set of existing tools, including Ansible, Terraform and OpenTofu, Helm, ArgoCD, Crossplane, Kubernetes Cluster API, CloudFormation, and Jenkins. Rather than replacing or refactoring these brownfield systems, the approach wraps each engine as a leaf-level activity within composite workflows, allowing cross-engine operations to be sequenced, monitored, retried, and recovered through a persistent workflow contract. The authors argue that this reduces reliance on fragile scripts and chained pipelines while preserving direct access to original tools for break-glass situations and enabling incremental adoption. Temporal Workers also support distributed execution through outbound, “inside-out” connections from the environments where tools run, which can simplify operation across secured networks. The same architecture could later incorporate AI agents and LLM-based tasks as additional workflow activities alongside deterministic infrastructure automation.

At the AI Engineer World's Fair, Temporal showcased an innovative and engaging booth featuring a massive 8-foot USB-C keyboard and holographic fan display, leveraging the Temporal Workflow for a unique Wordle game experience. The giant keyboard, conceptualized by Kristin Hillary and physically built by WhiteClouds, became a central attraction, as its interactive design required attendees to use it directly, revealing real-time Temporal Workflow history on a holographic fan. This setup highlighted how Temporal's durable Workflow architecture managed the game's state seamlessly, even accommodating live updates and user errors during the event. The booth demonstrated how a robust, central Workflow could enable flexible, engaging experiences while maintaining system integrity, with Temporal's architecture serving as the backbone for both the game and the surrounding interactive elements. Despite the logistical challenges, the project successfully illustrated Temporal's capabilities in handling long-running application states, even amidst the unusual constraints of a conference setting.

Crystal Palace has evolved from being a beloved underdog to a club competing in the Europa Conference League, marking its first foray into European football in 164 years. The club's new partnership with Temporal, a tech company known for ensuring task reliability across various industries, reflects a shared commitment to community and resilience. Temporal's software already underpins many everyday tasks, and its collaboration with Crystal Palace extends to enhancing fan experience and supporting local youth through computer science and AI workshops. This partnership is symbolized by the Temporal name on the club's shirt, alongside a return of the iconic sash design, as both entities look forward to a future of innovation and community engagement.

Temporal has launched the LangGraph integration for Python, currently in Public Preview, which allows developers to run agent frameworks like LangGraph on Temporal without altering their existing codebase. This integration offers automated failure recovery, human-in-the-loop capabilities, and robust handling of long-running processes. LangGraph simplifies agent definition but lacks comprehensive production features such as durable execution and recovery, which Temporal provides as an agent orchestrator. Temporal ensures durable execution where each node in a LangGraph can execute as a Temporal Activity, allowing processes to survive crashes and interruptions without manual intervention. Additionally, Temporal handles human-in-the-loop scenarios by allowing workflows to pause without consuming resources, automatically resuming once human input is received. The integration supports extensive observability through LangSmith, ensuring traceability across workflows and activities. Temporal's capabilities extend beyond LangGraph, supporting various frameworks and languages, and is designed to enhance the reliability and scalability of agent operations in production environments.

Unblocked, a context engine for AI-driven software development, transitioned its architecture to Temporal to address challenges in continuously computing against the current state of various data sources, such as code, PRs, and Slack conversations, without relying on cached answers. Originally utilizing a DIY actor model with custom queues and retries, Unblocked found this approach resource-intensive and opted for Temporal due to its polyglot SDK support, durable execution primitives, and robust capabilities. By integrating Temporal, Unblocked improved system reliability and efficiency, enabling deployment-free workflows and eliminating complex orchestration code. This shift allowed the engineering team to focus more on product development rather than infrastructure, significantly enhancing the overall value of the platform. With Temporal, Unblocked now efficiently runs about 1 million workflows daily, with plans to further streamline its code ingestion and expand Temporal's adoption across all engineering processes.

The narrative explores a lifelong fascination with the Japanese special effects genre, tokusatsu, which includes shows like Kyoryu Sentai Zyuranger, the original source for Mighty Morphin Power Rangers. This interest led the author to create Toku Solutions, a pipeline designed to translate Japanese toy manuals into editable static websites. This system uses OCR, translation APIs, and AI cleanup to overcome the challenges of translating complex manuals without losing the layout integrity essential for understanding. Despite technological aids, the process recognizes the limitations of AI, allowing for manual corrections to ensure accuracy. The pipeline is built using Temporal, a workflow orchestration tool that manages the various phases of OCR, translation, site generation, and AI cleanup, ensuring progress is not lost despite potential failures or retries. The entire endeavor is driven by the author's desire to fully understand and utilize collectible toys, with the code open-sourced for others who share the same interest.

Preeti reflects on her three-year journey with Temporal, highlighting her transition from HashiCorp to pursue challenging and rewarding problems in software development. She emphasizes her belief in Temporal's mission and the impressive talent and vision of its founders, Max and Samar. With the company's growth following its Series D funding and the integration of AI to enhance agility, Preeti is excited to lead as EVP of the Build, Operate, and Secure Organization, a role that combines research, development, and operational reliability. She aims to streamline the development process, reduce idea-to-feature time, and maintain product consistency while focusing on the reliable operation of AI systems. Preeti is grateful for the trust placed in her and is eager to continue solving complex problems with a passionate team.

Temporal and Encore are complementary tools that enhance backend operations by addressing distinct challenges: Temporal ensures the reliability of long-running processes by preserving workflow states, allowing them to resume seamlessly after interruptions without repeating completed steps, while Encore automates infrastructure provisioning across environments, streamlining deployment processes with integrated defaults and guardrails. These tools can be used together to create an efficient and robust backend system, as evidenced by companies like Dreem and Gradient Labs, who have integrated Temporal for workflow management and Encore for infrastructure automation, allowing for smooth operation and easy recovery from failures. The integration of these technologies provides a seamless experience where backend code is managed efficiently, from handling retries and replays in Temporal workflows to provisioning services in Encore, ultimately enabling software-first operations with a strong focus on reliability and operational clarity.

Rapidflare, founded by Dipkumar Patel and Vasanth Asokan, specializes in creating Agents for technical sales teams in the electronics and semiconductor industry, dealing with complex and voluminous product information. The company's focus is on developing an efficient document ingestion pipeline that can handle the intricate task of processing vast amounts of technical literature, ensuring high precision and reliability in responses to complex queries. The pipeline is designed to be durable, stateful, concurrency-controlled, observable, approval-gated, source-agnostic, and fresh to manage the dynamic nature of customer documentation. Rapidflare employs Temporal for orchestrating these processes, leveraging its features like durable execution, first-class Retry Policies, and sliding window mechanisms to manage the challenges of scaling, resource exhaustion, and failure management. The pipeline's architecture involves connecting, crawling, processing, and approving documents in a manner that minimizes downtime and maintains data integrity, with a focus on continuous improvement and adaptability to changing customer needs. The company's approach emphasizes using cloud storage for data transfer, implementing approval stages for quality assurance, and preferring Temporal's built-in scheduling over external solutions to enhance visibility and control in automated workflows.

Akshat Sandhaliya, CTO and co-founder of Sherlocks AI, has developed an AI-powered Site Reliability Engineering (SRE) platform that autonomously handles production incidents to reduce Mean Time to Resolution (MTTR) from hours to minutes. The platform deploys AI agents to investigate incidents by querying observability tools and tracing issues back to their root causes, thus mimicking a seasoned engineer's approach. The challenge lay not in developing the AI but in ensuring the reliability of the agents, which often run multiple tasks in parallel across several stages, including identification, investigation, root cause analysis, and remediation. Sherlocks AI overcame these challenges by using Temporal, which unifies four distinct workflow processes—Knowledge Graph construction, infrastructure scanning, AI agent investigations, and event ingestion—into a single reliable system. Temporal provides durable execution, retry policies, scheduling, and model agnosticism, allowing for seamless integration of various models based on task requirements. This approach eliminates operational overhead, enhances system reliability, and supports continuous agent improvements. The platform also incorporates human-in-the-loop mechanisms for complex cases and uses Temporal’s Scheduler for proactive anomaly detection, ensuring that agents can react to potential issues before they escalate into incidents.