What It Does

The three big LLM APIs do the same job with almost none of the same field names. Anthropic puts the system prompt at the top level; OpenAI puts it as the first message; Gemini calls it systemInstruction. Tools are input_schema here, parameters there, functionDeclarations somewhere else. Streaming, JSON output, and token limits each have three spellings.

This tool lets you configure a request once in a plain form and emits a correct, runnable snippet in curl, Python (official SDK), or TypeScript (official SDK) for whichever provider you pick. It is the reference you reach for when a parameter is being silently ignored.

What It Covers

Model, system prompt, user message, max tokens, temperature.
Streaming - with the correct per-SDK consumption loop (Anthropic’s event stream, OpenAI’s chunk deltas, Gemini’s stream iterator).
JSON output - response_format for OpenAI, responseMimeType for Gemini.
Tools - each provider’s function-definition shape, generated from one JSON array you paste.
Provider extras - Anthropic prompt caching (cache_control on the system block) and extended thinking (budget_tokens); OpenAI reasoning effort.

The Key Rule

There is no field for your API key, and there never will be. The snippets reference the standard environment variable ($ANTHROPIC_API_KEY, $OPENAI_API_KEY, $GEMINI_API_KEY), which the SDKs read automatically. Nothing you type is transmitted anywhere; the whole tool runs in your browser.

Pairs with the LLM Cost Calculator (what the request will cost) and the Prompt Cache Optimizer (whether caching is worth it).

Limitations

Chat-shaped requests. It builds single-turn chat/messages calls. Multi-turn histories, the OpenAI Responses API, batch, files, and the assistants/agents endpoints are out of scope.
A curated model list, verified on the date shown. Provider model names change; check the label date and swap the model string if a newer one shipped.
Snippets are a starting point. They are correct for the shapes above, but you should read your provider’s docs for the exact current parameter surface before shipping to production. Model IDs and parameter names move.