September 2025 Summaries

Inngest has introduced real-time support for Python, enabling developers to create interactive applications that deliver updates from durable workflows directly to the browser, overcoming the challenges of building reliable and interactive Python applications. Traditional setups using tools like Celery, RabbitMQ, and WebSockets often become complex and difficult to manage, especially when integrating additional solutions such as Redis Streams or Kafka. Inngest's new release simplifies this process by combining asynchronous workflows with real-time capabilities, offering managed WebSocket infrastructure to streamline development. This allows for the creation of rich user experiences, such as interactive AI applications, collaborative tools, and real-time updates for long-running processes. The latest Python SDK includes an experimental real-time package that supports publishing messages in workflows without the need for complex infrastructure management, focusing instead on business logic for faster, more reliable execution.

Replit has integrated Inngest into its Agent 3 to simplify the creation of durable, production-ready AI agents for developers, providing features such as retries, state management, and observability without requiring deep infrastructure knowledge. This integration allows users to interact with Replit's AI to develop applications by describing their needs, which the AI then codes, tests, and deploys. Previously, developers needed to manage complex tasks like state handling and tool integration manually, which was time-consuming and prone to errors. With Inngest, Replit users benefit from automatic retries, built-in observability, durable state management, rate limit handling, and graceful degradation, making it easier to manage and troubleshoot workflows. The integration also allows users to iterate on their agents using plain language, test them in a development environment, and deploy seamlessly to production with all the reliability of Inngest, greatly enhancing the user experience and efficiency in building robust AI-driven applications.

useAgent is a React hook designed to simplify the development of real-time, multi-step, and multi-agent backend workflows, enabling seamless integration with front-end user interfaces. It addresses three primary challenges in building such systems: ensuring durable workflows that can recover from failures, providing comprehensive observability for debugging complex agent behaviors, and streaming real-time experiences without extensive infrastructure management. Traditional approaches, like building custom solutions or using general agent frameworks such as LangGraph or Vercel AI SDK, often require significant setup time and lack complete features like durability and built-in observability. useAgent, built on Inngest's infrastructure, offers automatic retries and fault tolerance, composable streaming parts for real-time updates, and comprehensive debugging tools, all with minimal setup, allowing developers to focus on creating polished AI-powered workflows quickly.

Inngest has launched a beta version of its Insights feature, enabling users to interactively query and analyze event data directly from their workflow tools using SQL, without needing additional scripts or data pipelines. This tool addresses challenges faced by engineers in understanding the performance and cost of complex AI workflows by providing instant answers, unified context, and direct access to system, user, and usage data. Insights facilitates the extraction of actionable intelligence, making it easier for users to investigate, debug, and iterate on their processes. It is particularly valuable for AI-enabled products, where tracking token usage and model performance is crucial. Future updates will include the ability to create custom APIs and dashboards for more nuanced data visualization and analytics.

Step.Run introduces a solution for enhancing the durability of APIs by allowing developers to add a single line of code to their REST APIs, thereby incorporating automatic retries and comprehensive observability without the need for additional infrastructure. Traditionally, achieving API reliability required complex setups involving queues or custom background systems, but Step.Run simplifies this process by providing a straightforward method to maintain a single code path, akin to improvements seen in backend jobs. This approach ensures APIs can handle crashes, deploys, and cold starts while offering a clear view of every request, input, output, and error to facilitate quick troubleshooting and iteration. The platform is currently in developer preview for Go, with plans to expand support to TypeScript, enabling developers to experience these enhancements in real-time through a dashboard that tracks and manages API requests seamlessly.

Inngest has secured a $21 million Series A funding round led by Altimeter, with participation from A16z, Notable, Afore, and Guillermo Rauch, to advance its mission of enabling faster iteration in product development. The company offers a step-function architecture with built-in observability, allowing engineers to quickly productionize workflows and agents without dealing with infrastructure complexities. As the demand for AI-enabled products grows, Inngest aims to provide solutions that address the limitations of legacy systems and low-code tools by supporting durable, observable, asynchronous, and abstracted execution. This approach empowers engineers to manage complex workflows, enabling rapid iteration and adaptation to change, which is crucial in today's fast-paced development landscape. With plans to enhance its platform in areas such as API access, prototyping, and observability, Inngest is focused on supporting teams that thrive amidst uncertainty, helping them adapt swiftly to evolving technological demands.