Prompt Info

About the utility

Prompt inspection for model operators.

Prompt Info is a small technical workspace for measuring token count, projected cost, and payload format before model execution. It is built by Hello.World Consulting for developers, product teams, and AI engineers who need fast prompt checks without a login flow.

The tool is meant for preflight review: paste or draft a request, choose a tokenizer, inspect estimated usage, and compare structured payload shapes before a call reaches a provider. It keeps the practical planning details close to the prompt instead of burying them in separate spreadsheets or docs.

The product is intentionally narrow. It focuses on client-side measurement, transparent assumptions, and repeatable reference examples for teams evaluating model usage. That scope keeps the interface quick, reduces data handling surface area, and makes the results easier to verify during engineering review.

By Hello.World Consulting. .

Token planning

Paste a prompt, pick a model, change tokenizers, and estimate cost before sending.

Format comparison

Render the same prompt as TOON, JSON, YAML, XML, and CSV for practical payload review.

Resilient data path

Use a same-origin pricing API when available, with a static model catalog as the fallback.

Operational notes

  • Token counts use gpt-tokenizer, with selectable OpenAI BPE tokenizers: o200k, cl100k, p50k, p50k edit, and r50k.
  • Cost estimates are planning values. Provider billing may differ by model version and feature use.
  • The interface is designed for preflight review, not storage or user account management.
  • Format output is generated for inspection and planning, not as a guarantee that every downstream API accepts the same structure.
  • Model metadata is treated as a working catalog. When the live pricing endpoint is unavailable, the interface falls back to the bundled reference file so the utility remains usable.