01
LLM Token Counter
Estimates how many tokens a piece of text costs, broken down by script, and shows what…
System Prompt Linter
Checks a system prompt for the patterns that reliably cause trouble, and shows what it costs on every call.
Developer
Nothing uploaded
No sign-up
Free, unlimited
Paste something on the left to see the result.
In short
The two faults that most often break a system prompt in production are not wording problems. The first is never stating the output format — if a response is consumed by code, the prompt must say exactly what shape it must take, or the parser breaks on the tenth response rather than the first. The second is burying the critical constraint among a dozen instructions all marked "very important", which makes none of them important. Worth knowing too: the system prompt is resent on every single call, so its token count is a fixed cost per request and exactly the part that prompt caching is designed to make cheap.
How to use the System Prompt Linter
- Paste the prompt.
- Read the findings, then the token count.
- Test against known inputs after every edit — a linter cannot replace that.
About the System Prompt Linter
The System Prompt Linter checks a system prompt for the patterns that reliably cause trouble, and shows what it costs on every call. It runs as plain JavaScript inside the page, which means there is no upload step, no queue and no size limit imposed by a server — the only limit is your own device.
What people use it for
- Reviewing a prompt before it goes to production.
- Finding out what a long prompt costs per call.
- Giving a team a consistent checklist for prompt reviews.
Details
| Biggest fault | Not stating the output format |
|---|---|
| Second | Everything marked important, so nothing is |
| Cost | Paid on every call — cache it |
| Positive instructions | Followed more reliably than negative ones |
The token count is easy to overlook and worth taking seriously. A 2,000-token system prompt on a hundred thousand calls a month is two hundred million tokens of pure overhead — which is both a real bill and the ideal candidate for prompt caching.
Where this tool stops
Pattern checks on text. It cannot detect contradictions between instructions, judge whether the instructions suit the task, or predict how a particular model will follow them.
Frequently asked questions
Is "You are a helpful assistant" really a problem?
It is not harmful, it is empty. It occupies the most valuable position in the prompt — the opening, which gets the most attention — and says nothing the model does not already assume. Replacing it with what the model is doing, for whom, and what it should know is free and almost always an improvement.
Why prefer positive instructions to negative ones?
Because a negative instruction still puts the idea in the context. "Do not mention pricing" makes pricing salient; "answer only about features and refer pricing questions to sales" describes the behaviour you want. Both are understood, but the second is followed more consistently, and the difference grows with prompt length.
Does the linter tell me if my prompt works?
No, and nothing can except a test set. These are structural checks against patterns that reliably cause trouble. The only reliable way to know whether a prompt works is a set of inputs whose right answers you already know, run again after every change.
Does the System Prompt Linter cost anything?
There is no cost and no cap. The System Prompt Linter is not a demo of a paid product, so nothing is held back or watermarked.
Does anything I enter leave my computer?
No. The System Prompt Linter runs entirely in your browser using JavaScript. Nothing is transmitted, logged or stored, which you can verify yourself in the Network tab of your browser's developer tools.
Who is the System Prompt Linter for?
Typical uses: reviewing a prompt before it goes to production. Finding out what a long prompt costs per call. Giving a team a consistent checklist for prompt reviews. It is built for exactly those cases rather than being a general-purpose tool with this bolted on.
Is there anything the System Prompt Linter cannot do?
Pattern checks on text. It cannot detect contradictions between instructions, judge whether the instructions suit the task, or predict how a particular model will follow them. Knowing where a tool stops being trustworthy matters more than the cases where it works.
What figures does the System Prompt Linter work from?
Biggest fault is Not stating the output format. Second is Everything marked important, so nothing is. Cost is Paid on every call — cache it. These are printed on the page itself, so you do not have to take them on trust.
Will this work on a tablet?
Yes. The page is responsive and the utility works in Chrome, Safari, Firefox, Edge and Samsung Internet on both Android and iOS. There is no app to install and nothing to sign into.
Related developer tools
All 9202
Chat Transcript Splitter
Splits a long transcript into chunks under a token budget, breaking on speaker turns rather…
03
JSON Formatter
Formats, indents and validates JSON with precise error positions, key sorting and a…
04
JSON Validator
Checks whether JSON is valid and reports the exact line, column and reason for the first error…
05
JSON Minifier
Strips all whitespace from JSON to produce the smallest valid output, showing the bytes saved.
06
JSON to CSV Converter
Converts a JSON array of objects into CSV, flattening nested keys and handling missing fields…
07
CSV to JSON Converter
Turns CSV or TSV into a JSON array of objects, with type detection for numbers and booleans.
08
JSON to XML Converter
Converts JSON into well-formed XML with a configurable root element and array item naming.
09
XML to JSON Converter
Parses XML and converts it to JSON, preserving attributes and handling repeated elements as…
10
JSON to YAML Converter
Converts JSON to clean YAML with correct indentation, quoting and multi-line string handling.