Free AI tools
9 small tools for the parts of working with AI that are guesswork otherwise — what a prompt is missing, what a document costs to send, which model to use. No account, no limits, nothing kept.
Which one do you want?
Find the job, not the tool.
| If this is your problem | Use |
|---|---|
| I do not know why my AI output is disappointing | Prompt Scorer → |
| I want to know what this text will cost to send | Token Counter → |
| I have a task and want a structure to write it in | Framework Finder → |
| I want a decent Midjourney or DALL-E prompt | Image Prompt Builder → |
| I rewrote a prompt and want to know if it improved | Prompt A/B Compare → |
| I am building an assistant and need its instructions | System Prompt Builder → |
| My writing sounds like it was written by a machine | AI Writing Check → |
| I need to feed a document to an AI | Markdown Converter → |
| I do not know which model to use, or what it costs | Which AI Model? → |
Why these are free, honestly
"Free tool" usually means a trial, a credit balance or an email wall. These are none of those, and the reason is structural rather than generous: almost every tool here is ordinary code, not an AI call. Scoring a prompt, counting tokens, matching a task to a framework, converting a document — all of it is pattern matching and arithmetic. It costs us essentially nothing per use, so there is nothing to meter and no reason to ask who you are.
That is also the constraint that shaped them. A tool that called a model on every visit would need an account and a quota within a week, so anything that could not be done deterministically was either built a different way or not built at all. The paid product is the prompt generator, which does call models, and that is where the economics live.
Why the same input always gives the same answer
Ask an AI to score your prompt out of 100 twice and you will often get two different numbers. That is fine for a conversation and useless for a decision — if the score moves on its own, you cannot tell whether your edit helped.
Everything here is deterministic. The same text produces the same score, the same ranking and the same token count, every time and for everyone. When a number changes, something in your input changed. That is the whole reason comparing two prompts works at all.
The trade is that a deterministic checker cannot judge whether your idea is any good. It can tell you the audience is unstated, the format is unstated and there is no length target; it cannot tell you the piece is boring. Each tool says which of those it is doing.
What happens to what you paste
Nothing is stored and nothing is logged. Not the prompt, not the draft, not the document. These tools attract exactly the material people should be careful with — unpublished writing, internal documents, client work — and the only durable answer to that is to keep none of it.
The document converter goes a step further: your file never leaves the browser at all. Word, Excel, PowerPoint, PDF and the rest are parsed on your own machine in JavaScript, so there is no upload to trust. It is also why that tool can handle a contract.
Every file format the converter handles
The converter has a page of its own for each format, covering what that file actually records, what therefore survives conversion, and what cannot.
Questions
Are these AI tools really free?
Yes, with no account and no limit. Almost all of them are deterministic code rather than AI calls, so running one costs us nothing — which is what makes free sustainable rather than a trial. The two that do call a model are clearly marked.
Do you store what I paste into these tools?
No. Nothing you paste is stored or logged. The document converter goes further and never sends your file anywhere at all — it runs entirely inside your browser, which is why it can accept contracts and internal documents.
Why do the scoring tools always give the same answer?
Because they are deterministic. Ask an AI to score the same prompt twice and you will often get two different numbers, which makes it useless for comparing two drafts. These score the same text the same way every time, so a change in the number means a change in your text.
Which tool should I start with?
If your AI output is disappointing, start with the prompt scorer — it names what is missing. If you are worried about cost, start with the token counter. If you are choosing a model or a provider, start with the model picker.