About MakeTools

MakeTools is a small collection of free tools for people who build with large language models and who move data between systems: calculators for token cost, context windows, GPU memory and vector index size, a document chunker, converters for tool definitions, and two code-conversion utilities.

Why it exists

Many online utilities ask you to paste prompts, schemas or database definitions into a server you know nothing about, hide how a number was calculated, or quote prices without saying where they came from.MakeTools takes the opposite approach: each tool does one job, runs inside your browser, shows its formula, and states its limits.

Who runs it

MakeTools is operated by Prajyot Mulik, an individual based in Pune, Maharashtra, India. It is an independent project, not affiliated with Anthropic, OpenAI, Google, Microsoft or any other company named on the site. Those names appear only to describe the products the tools work with.

What is available today

LLM engineering

• LLM Token Cost Calculator – Tokens in, cost per call, per 1,000 calls and per month.
• Context Window Checker – Will your prompt, chunks and answer fit the context window?
• Prompt Caching Calculator – Cached vs uncached input cost, and the break-even point.
• Local LLM VRAM Calculator – Weights + KV cache + overhead, and whether it fits your GPU.
• Vector DB Size Calculator – RAM and disk for N vectors at a given precision.
• RAG Text Chunker – Split text by characters, tokens or sentences with overlap.
• Tool Schema Converter – OpenAI ⇄ Anthropic ⇄ Gemini ⇄ MCP tool definitions.
• Fine-tuning JSONL Validator – Line-by-line checks for chat-format training data.
• LLM Rate Limit Planner – Max requests per minute, concurrency and queue time.

Data & code conversion

• T-SQL to PostgreSQL DDL Converter – CREATE TABLE (T-SQL) → PostgreSQL DDL with a documented mapping.
• JSON to Pydantic Models – JSON sample → Pydantic v2 models.

Why everything runs in your browser

The tools are plain JavaScript that runs on your device. The site has no server code that receives your input: there is nothing to log, leak or sell. A side benefit is that, once a page has loaded, the tools also work offline. The one thing a few tools download is an open-source tokenizer table, from this site, when you ask for exact token counts.

How prices and model data are handled

Prices and model limits live in two data files. Every value records the page it came from (the provider’s own pricing or model page, or the model’s config.json on Hugging Face) and the date it was checked, and each page shows the oldest of those dates. A value that cannot be verified from an official source is left out rather than guessed. An automated check fails the build if any value has not been re-checked in 30 days. See the changelog for updates.

How the tools are tested

• All calculation and conversion logic is deterministic code, separate from the page, with automated unit tests for normal and edge cases (zero values, limits, malformed input).
• Every worked example on a tool page is produced by that same logic when the site is built, and the tests assert the same numbers, so an example cannot drift from the real behaviour.
• Browser tests exercise each tool, check accessibility with automated rules, and confirm the pages make no third-party requests and set no cookies.

Tests reduce mistakes but cannot guarantee there are none, and results are estimates where the page says so. If you find a problem, please tell us.

Advertising and independence

The site may display advertising in the future to cover running costs. If that happens, ads will be labelled “Advertisement”, kept apart from tool controls, and described in the Privacy Policy before they appear. Tools and guides are not written for, or changed by, advertisers.

Corrections and feedback

See the editorial policy for how guides are written and corrected, or contact us with suggestions.