API Tools

ArrayKit's API tools help developers, testers and anyone wiring up an AI model send requests, turn them into code and check what they cost. The API Request Client sends GET, POST, PUT, PATCH or DELETE requests with your params, headers and body, shows the status, timing and formatted response, and keeps your last 25 requests in this browser. For AI work, the OpenAI API Tester, Anthropic Claude API Tester and Google Gemini API Tester, plus testers for Replicate and Hugging Face, build a request from a model and prompt, run it where the provider allows browser calls, and copy it as cURL, Python or JavaScript. The Token Counter and LLM API Cost Calculator size a prompt before you pay for it. Twelve converters rewrite a curl command as code, among them cURL to Python, cURL to Fetch and cURL to PHP. For links, the URL Builder & UTM Generator adds UTM and custom parameters, and the URL Parser splits a URL into its parts. HAR to cURL and Postman to cURL turn a browser capture or a collection into curl commands, with an option to hide cookies and auth headers. Conversions and counts run on your device; only the API Request Client and the AI testers send anything, and only the request you built, straight from your browser to the endpoint you chose, never to ArrayKit.

  • cURL to Fetch — Convert a curl command to a JavaScript fetch() call.
  • URL Parser — Break a URL into its parts and list query parameters.
  • API Request Client — Send HTTP requests, build headers and params, inspect responses — your last 25 requests are saved locally.
  • URL Builder & UTM Generator — Construct and validate fully encoded production URLs with standard analytics query flags (UTM) or custom query arrays.
  • OpenAI API Tester — Build, run and copy OpenAI Chat Completions API requests as cURL, Python and JavaScript.
  • Anthropic Claude API Tester — Build, run and copy Anthropic Claude Messages API requests as cURL, Python and JavaScript.
  • Google Gemini API Tester — Build, run and copy Google Gemini generateContent API requests as cURL, Python and JavaScript.
  • Replicate API Tester — Build, run and copy Replicate prediction API requests as cURL, Python and JavaScript.
  • Hugging Face API Tester — Build, run and copy Hugging Face Inference API requests as cURL, Python and JavaScript.
  • cURL to Python — Convert a curl command to Python requests code.
  • cURL to Go — Convert a curl command to Go net/http code.
  • cURL to PHP — Convert a curl command to PHP cURL code.
  • cURL to Node.js (axios) — Convert a curl command to Node.js axios code.
  • Token Counter — Count tokens for GPT, Claude and Gemini live — exact OpenAI counts plus Claude/Gemini estimates.
  • LLM API Cost Calculator — Estimate GPT, Claude and Gemini API costs from token counts, with custom price overrides.
  • cURL to Rust — Convert a curl command to Rust reqwest code.
  • HAR to cURL — Turn requests from a browser HAR file into runnable cURL commands, with secrets stripped.
  • Postman to cURL — Turn a Postman collection export into runnable cURL commands, variables resolved.
  • cURL to C# — Convert a curl command to C# HttpClient code.
  • cURL to Java — Convert a curl command to Java 11 HttpClient code.
  • cURL to Ruby — Convert a curl command to Ruby Net::HTTP code.
  • cURL to PowerShell — Convert a curl command to PowerShell Invoke-RestMethod.
  • cURL to Swift — Turn a curl command into Swift code that uses URLSession.
  • cURL to Dart — Turn a curl command into Dart code for Flutter using package:http.

Guides

  • AI Tokens and Token Counts, Explained — What a token is in AI, how many words a token is, why GPT, Claude and Gemini count the same text differently, and how tokens set context limits and API costs.

Frequently asked questions

What can I do with these API tools?
Send and inspect HTTP requests with the API Request Client, try the OpenAI, Claude, Gemini, Replicate and Hugging Face APIs with your own key, and turn curl commands into code for languages from Python and Go to Swift and PowerShell. You can also convert HAR captures and Postman collections into curl, take apart or build URLs with UTM tags, and count tokens or estimate the cost of an LLM call before you make it.
Which languages can the cURL converters produce?
JavaScript with fetch, Node.js with axios, Python requests, Go net/http, PHP cURL, Rust reqwest, C# HttpClient, Java 11 HttpClient, Ruby Net::HTTP, PowerShell Invoke-RestMethod, Swift URLSession and Dart package:http. Each converter reads the method, headers, body, basic auth and -G query data the way curl does, and warns about flags such as -k that have no equivalent. To get a curl command in the first place, use Copy as cURL in your browser's developer tools, or convert a capture or collection with HAR to cURL or Postman to cURL.
Which of these API tools send data over the network?
Only the API Request Client and the five AI API testers, and only when you press Send or Run. The request you built then goes straight from your browser to the URL or provider you chose, and ArrayKit does not proxy, see or store it. Keys typed into the AI testers stay in browser memory, the client's history stays in this browser's local storage, and the converters, URL tools, Token Counter and cost calculator work entirely on your device.
How does the API Request Client compare with Postman?
It covers the everyday loop of building a request, sending it and reading the status, headers and formatted response, with nothing to install, and it can copy any request as cURL. Because it runs in a web page, it can only read responses from APIs that allow cross-origin (CORS) requests, and it keeps a history of your last 25 requests rather than collections and environments. When an API blocks browser calls, copy the request as cURL and run it in a terminal, where CORS does not apply.
How accurate are the token counts and cost estimates?
OpenAI counts are exact, because the Token Counter uses the same encodings as GPT-4o, GPT-4.1, the o-series, GPT-4 and GPT-3.5 Turbo. Claude and Gemini counts are labelled as estimates, since those providers do not offer an exact tokenizer that runs in the browser. The LLM API Cost Calculator multiplies your token counts by indicative prices that carry an as-of date, and lets you override the price per million input and output tokens for cached, batch or updated rates.