Tools

Tools you run in your browser

You put in your own numbers and the answer appears on the page. Nothing is stored and nothing reaches us.

Perform publishes two free tools that run entirely in your browser. The open role cost calculator shows what an unfilled technical role costs, measured against published hiring medians. The agent inventory shows how many of nine governance questions you can answer about the AI agents already running in your company. Neither asks for an email address.

Tools on this pageTwo
CostFree. No account, no email address, no form
Where the calculation runsIn your browser
What is sent to PerformNothing
Time per toolTwo to three minutes
Sources behind the defaultsUS Bureau of Labor Statistics, ManpowerGroup, Stack Overflow, Cloud Security Alliance

Do these tools store or send my numbers?

No. Both tools are HTML and JavaScript that run in your browser. The arithmetic happens on your device, nothing is written to a server, and Perform receives none of the figures you type.

Each tool has two share buttons and they do different things. Copy my results gives you your figures, the sources behind them, and a link that reopens the tool on that exact result, which is what you send to a CFO. Share the tool gives a clean link with none of your figures in it, so whoever opens it starts from the defaults and puts in their own.

Why is there no form?

Because a form would make the tool worth less to you and worth more to us, and that trade is the wrong way round. If a calculator is useful, it should be useful without a negotiation first.

If the output is worth a conversation, the contact page is one click away and a senior engineer reads what arrives.

Where do the default numbers come from?

Every default in both tools is published, cited on the page, and dated. Salary medians come from the US Bureau of Labor Statistics and the Stack Overflow developer survey. The hiring difficulty figure comes from the ManpowerGroup talent shortage survey of 39,063 employers. The agent governance benchmarks come from a Cloud Security Alliance survey of 445 practitioners.

Where a figure has no published source, the tool says so and does not guess. Four of the nine agent inventory questions have no benchmark anywhere, and the tool names them as unbenchmarked.

Start with a conversation

30 minutes with the engineer who would do the work.