The AI world moves fast, and we’re now entering the era of AI agents. These sophisticated AI systems integrate reasoning and memory, enabling them to autonomously interact with tools, make decisions, execute tasks, and deliver real-time results. AI agents will play an increasingly central role in how organizations operate, which Abhinav Asthana, our co-founder and CEO, discussed in a recent blog post. But systems this powerful are extremely challenging to build—especially without a unified solution. And the stakes couldn’t be higher: companies that are able to build and deploy agents gain a competitive edge, while those that lag behind risk obsolescence.
That’s why we’re incredibly excited to announce the launch of the Postman AI Agent Builder—a comprehensive suite of solutions that empowers developers to design, test, and deploy intelligent agents with ease. By combining LLM/API discovery and evaluation with visual workflow development, the AI Agent Builder removes complexity, allowing developers to focus on rapid innovation. Developers can build an agent from start to finish without leaving Postman, or they can build on their own platforms by using our discovery and tool generation APIs.
We’ve poured our hearts and souls into this launch, which will completely change the game of not just agentic AI, but of software as we know it. So, we’re dedicating the entire January edition of the Postman Drop to showing you what the AI Agent Builder can do. We’ll walk you through five key capabilities of the AI Agent Builder—whose features span the Postman API Network, API Client, and Flows—and demonstrate how you can start transforming your own organization with AI agents.
Evaluate APIs and LLMs with the world’s leading API client
Before you can build your agent, you need to know if an API or LLM meets your needs. This evaluation can be challenging and complex, especially because many APIs lack complete documentation, which makes it difficult to gauge how an API will perform under real-world conditions. Developers are left guessing about how an API or LLM works, which often leads to dead ends.
Postman has always been the world’s leading API client, enabling users to send requests, inspect responses, and debug any API from an intuitive interface. This lets developers easily explore APIs before making the decision to integrate them into their applications. Now, we’ve extended that same capability to LLMs.
When you find a LLM you’re interested in, you can use the AI protocol, integrated with the Postman API Client, to explore the outputs it gives to various inputs—just like you’re used to doing with APIs. The LLM Comparison Collection goes further, automating side-by-side comparisons of LLMs. It benchmarks key metrics like response time, token usage, and content quality—saving hours of manual work, enabling faster, data-driven decisions, and ultimately ensuring you choose the best LLM for the job at hand.
Whether you’re building AI-powered tools or applications, Postman helps you find the best model efficiently—so you can focus on building, not testing.
We have two video tutorials that can help you learn how to use these features. The first video walks you through the experience of sending requests with the new AI protocol, and the second video teaches you how to use the AI Agent Builder to evaluate and compare LLM performance. You can also get hands-on with these capabilities in our dedicated workspace.
Build agents and workflows without writing custom code
AI agents need stateful workflows with memory support to adapt to earlier interactions, as well as reinforcement learning and real-time data streaming to handle new information. But without the AI Agent Builder, putting the pieces together involves writing, debugging, and testing complex custom code. It’s enough to overwhelm even the most seasoned developer.
What’s a dev to do, you might ask? Enter: Postman Flows. Flows is a powerful visual tool that simplifies the development and sharing of API-powered applications—and now, AI agents. Developers can use its intuitive drag-and-drop interface to build powerful AI agents by connecting AI requests and API requests on a powerful visual canvas without writing code. They can also use Modular Flows to rapidly prototype and build multi-step agentic workflows—and leverage ready-to-use Flows templates with preconfigured connectors for common workflows.
And if you’d prefer to work in your own development environment, you can leverage our new Tool Generation API, which enables you to generate tools to integrate with APIs from the 18,000 companies on our network.
Check out our video tutorial, which shows you how to build an autonomous AI agent for incident management that analyzes Slack conversations, creates documentation, and makes critical decisions. You can also get hands-on with these capabilities in our dedicated workspace.
Test agentic solutions easily and holistically
Once you’ve built your agent with Flows, it’s time to test its behavior across a range of real-world scenarios. Testing software is always complex and time-consuming, but the stakes are even higher in the world of AI agents. Developers must ensure the agent aligns with their organization’s goals and values while protecting sensitive data and ensuring GDPR compliance. And because APIs don’t offer accessible, low-risk testing environments, they are often forced to experiment in production-like setups.
Postman Flows lets developers locally test their agentic workflows without the complexity of managing test infrastructure. You can also easily configure inputs, parameters, and variables to simulate different scenarios the agent might encounter. Users expect flawless performance from production-grade AI systems, and testing your agent with Flows helps ensure it consistently meets those high standards.
Get hands-on with these capabilities in our dedicated workspace.
Deploy and monitor production-ready agents
An agent that is built, tested, and fine-tuned is ready to deploy to production. This process involves integrating the agent with existing systems, ensuring consistent, low-latency performance, and leveraging AI-specific hardware and cloud platforms. This can be tricky, as compatibility issues, limited server capacity, and the complexity of large datasets make it difficult to maintain real-time performance and functionality.
Flows already allows agentic AI workflows to be developed and run locally, which enhances security and preserves data sovereignty. In the coming months, we will add support for deployment to the Postman cloud, which will provide the necessarily scalability and flexibility without the need to provision and manage the infrastructure. We will also enable an insights dashboard and cloud run monitoring, so that teams can keep tabs on performance metrics and ensure their live AI agent continuously meets requirements at scale.
We’re working hard on these capabilities and are excited to share them with you soon. Join the waitlist to be the first to know when cloud deployment and monitoring are available.
Discover APIs and LLMs from a centralized platform
In the era of AI, building agents is not one-and-done. We’ve established how essential APIs and LLMs are to building AI agents: LLMs enable humans to interact with agents, and APIs enable agents to interact with applications, retrieve information, and perform tasks. But when building an agent, developers often struggle to find APIs and LLMs that are relevant and reliable. To make matters worse, APIs often have multiple versions, which makes it challenging to know which version to use or if dependencies will break.
The Postman API Network plays a crucial role in helping developers get over this initial hurdle. It is home to 100,000 APIs from 18,000 companies—including Salesforce, PayPal, and UPS—making it a centralized source for instantly discovering verified and well-maintained APIs and LLMs.
Developers can easily find APIs, LLMs, teams, workspaces, and collections from within the API Network. We’ve also published a new Network Search API, which allows users—as well as AI agents—to programmatically search the network from any external environment. Agents can use this API to search the network as part of their workflows, which unlocks countless use cases and dramatically improves agent capability. Request access to the Network Search API here.
Get hands-on with these capabilities in our dedicated workspace.
Enter the new AI frontier
This special edition of the Postman Drop showcased how the new AI Agent Builder unifies LLMs, APIs, and workflows to streamline agent development—across every stage of the process. We are in the midst of a seismic shift in how APIs and LLMs are used, and we’re excited to offer this powerful solution for building AI agents within the same platform that 35 million developers already know and trust. We will continue to iterate and expand upon the AI Agent Builder’s capabilities, solidifying Postman’s position as the world’s leading API platform and the de facto place to build AI agents.
Explore the AI Agent Builder workspace to start working and experimenting with these new capabilities, and keep your eyes peeled for more updates.
Last but not least, please leave any feedback or questions in the comments! Our product team is eager to hear from you.
Keep it 200,
The Postman Team
tell us what you think.
I’m investigating creating video content for work, so this is great inspo.
. I’ve been digging into scripts recently and I’m wondering if there’s any we can ship with our collections to help make things easier for our customers — especially during precommittment.