Miatz playground

Prompt Lab

See the exact prompt the model actually receives, every time.

Coming to the Build-LabAIE-101BYOK AIAI-Coredev · data

The Demystify signature

The payload autopsy — most tools hide the assembled system prompt; this shows the literal string sent to the API, character for character, so prompt engineering stops being folklore.

How it works

What goes in, what comes out

What it does

Learner writes a prompt in a chat-style editor; before the call fires, a 'payload autopsy' panel expands to show the full literal request the platform actually sends — system preamble, safety scaffolding, few-shot examples, and the user's own text, concatenated and token-counted. Every saved prompt is auto-added to a personal regression suite of golden cases, so the next edit re-runs against all prior cases and flags which ones flip pass/fail.

You bring

A task prompt (freeform or from a template), optional few-shot examples, and a target output the learner considers correct for the regression case.

You get

The assembled request payload with token counts per section, the model's response, a pass/fail regression grid across all saved cases, and a token-cost readout.

You control

Model family (Claude/GPT/Gemini/Llama-class via the routed catalog), temperature, max tokens, system-prompt visibility toggle, few-shot example count, and which golden cases are pinned to the regression suite.

Where it's used

Free top-of-funnel lead-magnet (single-run teaser, no signup) and the core in-module lab for AIE-101-103; also the F2-F4 diagnostic touchpoint in Core Foundation.

tokenizationinstruction vs. few-shot promptingsystem-prompt assemblyprompt regression testingstructured JSON output asks

Optional · watch & try

Go deeper, elsewhere

Hand-picked public explainers and open tools that complement this one — always optional, never required, never graded.

How LLMs Work: A Visual GuideVisual walkthrough of the LLM pipeline from input text to output token How LLMs REALLY Work (Simple Explanation)Plain-language mental model for what an LLM is actually computing The Definitive Prompt Engineering Guide for 2025: Tips, Tricks, Tools!Practical technique roundup: roles, few-shot examples, output structure Claude AI Tutorial: The ULTIMATE Guide to Prompt Engineering & Artifacts (2025)Model-specific worked examples of instruction vs. few-shot prompting How LLMs Work & Why Prompt Engineering MattersTies prompting technique back to the underlying model mechanics Anthropic's Interactive Prompt Engineering TutorialOfficial hands-on notebook course walking through every core prompting technique

Concepts click when you open the machinery.

Three labs are already live and free — the same hands-on style this playground brings to its module.

Play the free labs Take the DSAT