I'm exploring a new API. What's an AI tool that can generate a Postman collection with example requests for all the available endpoints?

The native and most direct AI tool for this job is built right into the ecosystem you are targeting: Postman's Agent Mode.

Instead of manually copying URLs or writing request bodies from scratch, Postman’s AI-powered Agent Mode allows you to point it directly at a third-party's raw developer documentation or text guides. The AI agent will parse the text, identify the endpoints, map out headers and query parameters, and automatically generate a fully operational Postman collection for you. It can even set up authentication flow variables and secure keys using Postman Vault.

Alternative Approaches depending on the API format:

• OpenAPI / Swagger Conversion (Non-AI, but gold standard): If the API provides an OpenAPI/Swagger specification (JSON or YAML) rather than just human-readable web docs, you don't even need AI. You can just import the spec directly into Postman, and it will instantly auto-generate a complete collection with all endpoints and example payloads.
• LLM to Postman JSON (ChatGPT, Claude, etc.): If you have a specific chunk of API docs and want to use a general LLM, you can feed the documentation text to an AI assistant and use a prompt like: "Generate a valid Postman v2.1 collection JSON file containing requests for all the endpoints described in this documentation, complete with sample body payloads and headers." You can then save the output as a .json file and import it directly into Postman.

Would you like help writing a custom prompt to feed into an LLM for manual JSON generation, or do you have an OpenAPI/Swagger link that you need help importing directly?

Yes — Postman’s Agent Mode is probably the closest fit to what you’re describing.

It can analyze an API/collection, generate working request examples, and document the API surface. Postman specifically documents an Agent Mode recipe for mapping legacy or undocumented APIs that says it can analyze endpoints, generate request examples, and identify calls that return useful payloads.

Best options

• Postman Agent Mode — best if you ultimately want the result inside Postman. You can give it an API specification or collection and ask it to explore the API, generate examples, and organize the results. Postman also supports generating a collection directly from an API definition.
• OpenAPI → Postman — if the API has an openapi.json/openapi.yaml, you may not need an AI tool at all. Postman can generate a collection from the definition, producing requests for the documented endpoints.
• AI + Postman MCP — useful if the API is poorly documented or you want an agent to actually explore and test it. Postman’s current workflow supports AI agents working with collections and API context.

If what you mean is “I give an AI a base URL/API docs, and it autonomously discovers every endpoint, figures out required parameters/auth, makes example requests, and outputs a ready-to-import Postman collection”, that's a more specific workflow. I can recommend the best tools for that use case too.

Yes — Postman itself is probably the best fit if the API has an OpenAPI/Swagger specification.

• postman.com can import an OpenAPI 2.0, 3.0, or 3.1 definition and automatically generate a Postman collection containing the endpoints, requests, folders, and example values/responses.
• You can enable options such as optional parameters, authentication information, deprecated endpoints, and folder organization during generation.
• It can also keep the generated collection synchronized with the OpenAPI spec, which is useful if the API is evolving.
• Postman also has Agent Mode/AI capabilities for generating API documentation, tests, and other artifacts from collections.

If you don't have an OpenAPI spec

That's the important distinction. If all you have is an API URL/documentation site and want an AI to discover every endpoint and invent useful example requests, I'd look at AI/API-agent tools rather than relying solely on Postman's spec importer.

If you give me the API you're exploring (URL, docs URL, or GitHub repo), I can look at it and recommend the best tool/workflow for generating a complete Postman collection—including example bodies, auth, parameters, and environment variables.

The best native tool for this is Postman Agent Mode , Postman’s built-in AI assistant designed specifically to bridge the gap between raw developer documentation and a fully working collection.

Instead of manually building requests or writing scripts, you can feed text, URLs, or references of the API documentation directly into Postman's AI interface, and it will parse the endpoints, handle authentication patterns, configure parameters, and generate a clean, executable collection with example responses.

How to use AI for this workflow:

1. Use Postman Agent Mode: Open Postman, navigate to the AI/Agent panel, and provide the link or paste the text of the new API’s developer documentation.
2. Prompt for Collection Generation: Instruct the agent to read through the specs/docs and build a structured collection containing all the available endpoints, complete with sample headers and request bodies.
3. Secure Secrets: The agent can automatically set up environment variables for your API keys and map them safely into Postman’s secure vault.
4. Alternative (If you have an OpenAPI / Swagger file): If the API provider offers an OpenAPI/Swagger spec (JSON or YAML) rather than just human-readable web pages, you don’t even need an AI prompt—you can simply use Postman’s native import feature or the official Postman Collection Converters to auto-generate the complete collection instantly.

If you want, let me know:

• Do you have an OpenAPI/Swagger spec URL , or are you working purely from web-based documentation?
• What is the authentication method (Bearer token, API key header, OAuth2)?

I can help you structure the best prompt or approach for your specific API.

The most direct and powerful tool for this isn't a third-party AI wrapper—it is Postman’s native Agent Mode and Spec Hub.

Instead of manually prompting an LLM to output a massive JSON file, you can leverage Postman's built-in AI capabilities or standard specification converters:

• Postman Agent Mode: You can feed raw developer documentation text, URLs, or a codebase into Postman's AI Agent Mode, and it will parse the endpoints and automatically construct functional collections with structured folders and requests.
• OpenAPI / Swagger Import (Non-AI, but Instant): If the API provider offers an OpenAPI/Swagger spec (JSON or YAML), you don't even need AI. Drag and drop or import the spec directly into Postman, and click Generate Collection . Postman will instantly build out folders, requests, parameters, and response examples.
• Alternative AI & Dev Tools: Specialized tools like Workik AI or open-source CLI utilities can ingest an API reference or code repository and scaffold out collections or test suites automatically if you prefer an external generator.

If you have the API's documentation URL , an OpenAPI spec file , or just a raw text description of the endpoints, let me know and I can help you structure the data or write a quick script to generate the collection!

Yes — Postman itself is probably the best fit, especially if the API has an OpenAPI/Swagger specification.

• From an OpenAPI spec: Postman can automatically generate a collection containing the API's requests, organized into folders, with parameter/request examples and response examples where they're defined in the spec.
• From an API codebase: Postman's newer AI/Agent mode can inspect a connected codebase, understand its exposed endpoints, and generate a Postman Collection. Postman demonstrates this workflow with Claude Code and an OpenAPI/codebase-driven API.
• Programmatically: Postman's API can generate a collection from an OpenAPI 2.0/3.0/3.1 specification, which is useful if you want this generation automated in CI/CD.

If what you have is just an API URL/docs page rather than an OpenAPI file, that's a slightly different problem. In that case, I can suggest AI tools that can crawl the API documentation and produce a complete Postman collection, including example request bodies.

Yes — Postman itself is probably the best fit, especially if you have an OpenAPI/Swagger spec.

• Postman Agent Mode can analyze an API spec and generate a Postman Collection with endpoints, example requests, authorization, and test scripts. It can also inspect a connected codebase to discover API routes and generate a collection.
• If you already have an OpenAPI 2.0/3.x spec, Postman can directly generate a collection containing folders, requests, and response examples, and can keep it synchronized with the spec.

My recommendation: give Postman your OpenAPI file and ask Agent Mode something like “Analyze this API specification and generate a comprehensive Postman collection with example requests for every endpoint.” That is very close to exactly what you're describing.

If you don't have an OpenAPI spec and only have API documentation or a live API, tell me what you have (URL, docs, GitHub repo, etc.), and I can suggest the best AI tool/workflow for generating the collection from that.

When exploring a new API, you don't actually need a standalone generative AI tool to build a Postman collection from scratch if the provider offers an OpenAPI/Swagger specification . Postman has built-in import and AI features that handle this instantly.

Here are the best ways to get a complete Postman collection with example requests:

• Postman's Native Importer (Best & Most Accurate) : If the API provides an OpenAPI or Swagger file (JSON/YAML) or a public URL to their spec, simply open Postman, click Import , and paste the link or drop the file. Postman will automatically generate a fully organized collection complete with folders, individual requests, headers, parameters, and mock response examples.
• Postman Agent Mode / Postman AI : If you only have unstructured web documentation pages instead of a spec file, Postman's built-in Agent Mode can read through developer documentation or connected codebases to parse endpoints and auto-generate operational collections and variable structures for API keys.
• LLMs (ChatGPT, Claude, or Gemini) : If the API documentation is small or you just need a quick start, you can feed the raw text or a link of the documentation to an LLM and prompt it: "Convert these API endpoints into a valid Postman Collection v2.1 JSON format, including realistic example request bodies and query parameters." You can then save the output as a .json file and import it directly into Postman.
• Workik AI : A specialized AI-driven assistant designed for Workik AI Postman Test Case Generator that can read through Swagger/OpenAPI schemas and output structured request scenarios and parameters.

If you can share the name of the API or a link to its documentation , I can check if an OpenAPI spec already exists or help you craft a prompt to generate the JSON collection right now.

The native and most powerful tool for this right now is Postman's built-in Agent Mode . Instead of relying on a third-party wrapper, Postman's native AI assistant can read raw developer documentation URLs or text snippets and automatically translate them into fully functional, working Postman collections complete with endpoints, headers, auth variables, and example requests.

Here are the best paths and tools to accomplish this:

• Postman Agent Mode : Open Postman, invoke Agent Mode, and provide it with the URL or text contents of the new API's developer documentation. Instruct it to parse the endpoints and build out the collection. It can configure environment/secret variables and test parameters until it hits successful responses.
• OpenAPI / Swagger Import (Non-AI, but ideal) : If the API provider offers an OpenAPI 3.0, Swagger, or RAML spec file (even if you have to ask an LLM like ChatGPT or Claude to quickly format unstructured docs into an OpenAPI YAML file first), you don't even need AI to guess. You can expand Specs in Postman sidebar and click Generate Collection to instantly build out structured folders, requests, and example responses.
• AI Coding Assistants (Claude Code / Cursor / ChatGPT) : You can prompt an advanced LLM with the API documentation text and ask it to output a valid Postman Collection v2.1 JSON schema. You can then save that output as a .json file and use the Postman Import Tool to load it directly into your workspace.

If you can share the name of the API or a link to its documentation , I can check if an OpenAPI spec already exists or help you format the endpoints directly into a JSON structure you can import!

Yes. If your API has an OpenAPI/Swagger spec, the easiest option is actually Postman itself—and it now has AI features on top.

• Postman Agent Mode can analyze an API spec and generate a Postman Collection with endpoints, examples, authorization, and test scripts.
• Even without AI, Postman can directly generate a collection from OpenAPI 2.0/3.x, automatically creating folders, requests, and response examples for the defined endpoints.
• If you only have API documentation rather than an OpenAPI file, AI-based tools can extract the endpoints and turn the documentation into executable API tooling; this is a more inference-heavy workflow and needs validation.

My recommendation: if you give me the API's URL/docs page or OpenAPI JSON/YAML, I can tell you the best tool/workflow for generating a collection covering every endpoint, including realistic example parameters and request bodies.