---
title: Discovery Is Now Nine One-Click Tools
url: https://varstatt.com/jurij/p/discovery-is-now-nine-one-click-tools
author: Jurij Tokarski
date: 2026-10-10
description: Eight AI chats became nine short forms, each with one AI call and a one-page result. A privacy question in another chat started it; session data finished it.
section: Blog (https://varstatt.com/jurij/archive)
tags: ai (https://varstatt.com/jurij/c/ai), project-stories (https://varstatt.com/jurij/c/project-stories)
---

A visitor in my [404 page's AI chat](/jurij/p/404-pages-as-lead-magnets) asked where the privacy policy was. They worried that a free-text chat might be storing private or personal data.

Fair question. It kept what visitors typed.

That question started the change that turned [Discovery](/discovery), my set of free tools for software founders, from chats into one-click tools.

## The First Split Kept the Chat

Discovery had been split once before. In March I wrote that [nobody finishes a 15-minute AI interview](/jurij/p/nobody-finishes-a-15-minute-ai-interview), and split it into eight tools, one deliverable each.

My plan was a shorter interview, with some of my own knowledge built in so the tools would argue back.

So each tool was still a chat. Chat is still the native format for AI products, and the first version had been one too.

## Every Form Was a Privacy Issue

The 404 question made me look at every place on the site that took free text from strangers. Each one was a privacy risk, whether anyone used it or not.

The 404 chat became a plain page. The principles chat went too; almost nobody had used it. That left Discovery as the last chat standing.

At first I only thought Discovery probably needed another iteration. Then the other signals lined up.

## Nobody Finished the Shorter Chats

The first signal was the data. The sessions stored from May to September told the story: none went past its first tool, and most stopped four or six messages in.

Shorter chats lost founders the same way the long interview had.

## Founders Came for the Result

The second signal was in the code. The March post said the deliverable is the unit, and the code agreed: about 2,900 of the 12,200 lines in Discovery existed only for the interview. The renderers, the PDF and the email already worked on finished results.

A founder who came for a lean canvas still had to talk their way to it.

## One Call Can Be Checked

So my brief on October 5 had three steps: the user provides data, one AI call, a formatted and structured result.

One call returns one structured answer, and code can test, cap and price it before anyone sees it. A chat's answer depends on every turn before it.

## Nine Tools, One Form Each

Discovery v5 went live on October 7 with nine tools: the eight, plus a new Idea Generator.

<PostImage
  src="/jurij/discovery-is-now-nine-one-click-tools-001.webp"
  caption="The hub: nine tools, one card each."
/>

Each tool starts with one short form: every question on one screen, one run button.

<PostImage
  src="/jurij/discovery-is-now-nine-one-click-tools-002.webp"
  caption="Lean Canvas asks five questions. Two are optional."
/>

The run is one Claude Sonnet call that returns structured output.

<PostImage
  src="/jurij/discovery-is-now-nine-one-click-tools-003.webp"
  caption="One call while the founder waits: answers, canvas, check."
/>

Code renders that output as a one-page result, the same on screen and in the PDF, with a Markdown copy.

<PostImage
  src="/jurij/discovery-is-now-nine-one-click-tools-004.webp"
  caption="The full result: a verdict first, then where the idea is weakest, then the canvas itself."
/>

Answers go to the model for the run and are not stored. Code also guards what comes back: if the model adds a sales line, like my name or "hire me", the run fails and shows the sample instead.

## A Test That Fills the Form Like a Founder

Guards need proof, so every tool has an end-to-end test. It opens the tool in a real Chrome, fills the form like a visitor, runs it and judges the result.

It screenshots every step and downloads the PDF, JSON and Markdown. Then it checks what the model wrote: caps, quotes, banned phrases, English only.

On October 10 it ran all nine tools against real model calls, around one made-up idea: a booking link for mobile dog groomers. A result took 9 to 24 seconds and cost about 4 cents.

<PostImage
  src="/jurij/discovery-is-now-nine-one-click-tools-005.webp"
  caption="One idea through all nine tools: nine one-page PDFs from real model calls."
/>

<PostImage
  src="/jurij/discovery-is-now-nine-one-click-tools-006.webp"
  caption="Nine live runs, no failed check."
/>

The loading and result screenshots above come from those runs too.

## What It Doesn't Do Yet

The tools don't chain yet: each one stands alone for now. It's also too early for a usage verdict, since search data for the new pages comes in over the next weeks.

## The Second Split Removed the Chat

The March post said to build around the deliverable. I did it for the pages and kept the conversation.

A privacy question took the conversation out, and the data said nobody would miss it.

If the visitor came for a document, a form gets them there in one click.

> Try it: [varstatt.com/discovery](https://varstatt.com/discovery)