When we build Agentic workflows, we find it tedious and time-consuming to integrate and test APIs.
For every API, we have to read the docs, understand endpoints, manage authentication, ensure all arguments are available to the API, and whatnot.
Recently, Postman released the AI Agent Builder—a suite of tools to streamline agent development by helping developers integrate any of 100,000+ APIs on the Postman network in less than a minute.
Let’s understand them below.
You can test these updates by creating an account here.
Update 1) LLM testing and evaluation
We have so many LLM providers now. Each provider has different authentication methods, different parameters, and more.
This makes it challenging to decide which provider to proceed with, the associated costs, etc.
This update gives you a nice interface to test prompts, responses, and performance of multiple LLMs simultaneously (along with respective settings):
You can also compare the performance of several LLMs using their Collection Runner or Flows.
This is depicted below:
While both provide immediate feedback, the key difference is that:
Collection Runner is perfect for large-scale testing.
Flows is ideal for interactive evaluation and parameterization to quickly test across different scenarios.
Update 2) Agent and workflow orchestration
This update gives you a visual, no-code canvas to quickly build, configure, and test multi-step AI workflows.
To understand better, imagine this customer support workflow:
It accepts the user query.
Retrieves relevant context (from past conversations, docs, etc.)
Decide if the context is relevant (this is a conditional):
If yes, it produces a response and returns it to the user.
If not, it notifies the human support team.
As shown in the video below, you can configure these workflows inside Postman’s AI Agent Builder and test all integrations to ensure they run as expected:
Update 3) Tool generation [Huge time saver]
Agents need tools to take real-world action.
This update lets you instantly build and test tools for your agents!
More specifically, Postman’s AI Agent Builder has a Tool Generation API, which makes it incredibly easy to turn any API into an agent-ready tool in three clicks.
The demo video depicts how to use this:
Find an API in Postman’s Network.
Pick a programming language and agent framework.
Generate production code that you can directly integrate into your AI Agent.
Done!
Conclusion
Orchestrating and testing agents (or any multi-step AI workflows) is tough.
We personally find it extremely time-consuming and tedious to integrate and test APIs in RAG and Agentic workflows that we showcase in this newsletter.
As mentioned above, for every API, we have to read the docs, understand endpoints, manage authentication, ensure all arguments are available to the API, and whatnot.
In this respect, Postman’s AI Agent Builder will be a huge time saver.
This way, AI developers can only do what they are supposed to—build agentic workflows and improve them without worrying about integration and testing challenges.
We’ll do a hands-on demo of using the AI agent builder in the coming weeks.
Postman’s AI Agent Builder is solving a big problem, and we are eager to see what more updates they will ship next.
Thanks to Postman for giving us access to their AI agent builder and partnering with us on today’s newsletter.
👉 Over to you: Do you find these updates exciting?
Thanks for reading!
Exciting read. Thanks.
This seems very promising. Looking forward to giving it a ride.