Stop paying AI tools for bad questions
SnapPrice rewrites vague prompts before they reach ChatGPT, Claude, Gemini, Copilot or Perplexity — so your team gets a usable answer on the first try, with fewer tokens and less waiting.
- 1Sign up as your team's adminFree for the pilot. Takes 30 seconds.
- 2Generate an API key for each team memberIn your admin dashboard. Send it by email or Teams. To activate the Chrome extension you have to sign up as admin to generate the API key.
- 3They paste it into the Chrome extensionSnapPrice now improves every prompt they type — and you see the savings.
How to install and use SnapPrice
SnapPrice works inside ChatGPT, Claude, Gemini, Copilot and Perplexity — in Chrome, Edge, Brave or Arc on a computer.
Install it (about 2 minutes)
- Download the extension (a small .zip file).
- Unzip it. You get a folder called
snapprice-extension. - In Chrome, open
chrome://extensions(Edge:edge://extensions). - Turn on Developer mode (switch at the top right).
- Click Load unpacked and choose the
snapprice-extensionfolder. - Click the puzzle icon 🧩 in the toolbar and pin SnapPrice so the icon is always visible.
- Click the SnapPrice icon, paste your API key (generated in Admin dashboard → Team & keys) and click Connect. It shows “Connected as your name”.
Use it every day
- Open ChatGPT, Claude, Gemini, Copilot or Perplexity and type your prompt as usual (5 words or more).
- Pause for a second. SnapPrice shows a sharper prompt, the techniques it used and the tokens and $ you save.
- Press Tab to use it, or Esc to keep yours. You can edit the suggestion before using it.
- Prefer to choose? Click ✨ SnapPrice or press Alt+S any time. In the popup you can switch to Fully automatic (with Undo) or Only when I click.
- Type in Japanese and you get a Japanese prompt. Your total tokens and money saved are in the popup.
Stop paying AI tools for bad questions
Type a prompt the way you normally would. SnapPrice rewrites it, runs both versions on the same AI model, and shows you the real difference in tokens and time. Works in English and Japanese.
About you & this session — optional, makes rewrites fit your real work
Your savings
Stored on this device onlyWhat this means for a company
Uses your measured reduction, or the 40.4% benchmarkAssumptions from the SnapPrice benchmark (1,000 prompts, 5 AI tools): 73.0s → 24.2s average time to a usable answer, 20 working days per month, ~1,200 tokens per unoptimized task. Time saved is scaled by the reduction you choose.
All testers
Your recent prompts
| Prompt | Optimized | Tokens | Time |
|---|---|---|---|
| No prompts yet — try one on the first tab. | |||
SnapPrice inside the AI tools you already use
Works inside ChatGPT, Claude, Gemini, Copilot and Perplexity. Just type — when you pause, SnapPrice suggests a sharper prompt with the expected tokens and cost saved. Press Tab to use it, or turn on fully automatic mode.
Get it running
It's already connected to this server. You only need your personal API key.
Install it (about 2 minutes)
- Download the extension (a small .zip file).
- Unzip it. You get a folder called
snapprice-extension. - In Chrome, open
chrome://extensions(Edge:edge://extensions). - Turn on Developer mode (switch at the top right).
- Click Load unpacked and choose the
snapprice-extensionfolder. - Click the puzzle icon 🧩 in the toolbar and pin SnapPrice so the icon is always visible.
- Click the SnapPrice icon, paste your API key (generated in Admin dashboard → Team & keys) and click Connect. It shows “Connected as your name”.
Use it every day
- Open ChatGPT, Claude, Gemini, Copilot or Perplexity and type your prompt as usual (5 words or more).
- Pause for a second. SnapPrice shows a sharper prompt, the techniques it used and the tokens and $ you save.
- Press Tab to use it, or Esc to keep yours. You can edit the suggestion before using it.
- Prefer to choose? Click ✨ SnapPrice or press Alt+S any time. In the popup you can switch to Fully automatic (with Undo) or Only when I click.
- Type in Japanese and you get a Japanese prompt. Your total tokens and money saved are in the popup.
What you see
One invisible layer. Every prompt, fixed.
Vague prompts make AI tools ramble and need retries. SnapPrice fixes the question before it's asked.
~40% fewer tokens
A specific prompt with a length and format gets a focused answer instead of an essay that covers every angle.
3× faster to a usable answer
Retries fell from 2.42 to 1.11 per task in our 1,000-prompt benchmark — that's where most of the time goes.
+82.5% prompting skill
Seeing a better version of your own prompt every day teaches by example — no training course needed.
About this prototype
SnapPrice keeps the context of each session — your optional background plus the prompts and answers so far — so a follow-up like “make it shorter” is rewritten about that answer, not something random. Write in Japanese and you get a Japanese rewrite and answer; write in English and you get English. Each comparison runs your original prompt and the SnapPrice version on the same AI model and reports the real token counts and response times. A single run measures one answer; the benchmark numbers above also count retries, so savings in a single run can be smaller or larger. Prompts are processed only to produce the result and are not used for AI training. Please don't paste confidential information.