Know what AI thinks.
And why.
Every week Tonecast asks ChatGPT, Claude, Gemini and Perplexity the questions your customers ask, shows what they say about your brand — and traces the web pages that shaped those answers.
The demo is public and needs no account. It runs on fictitious brands.
“Aurora Coffee is a solid choice for compostable capsules: the redesigned capsule passed an independent jam test [1] … prices start at €0.35 per capsule [2].”
Shaped by: brewjournal.example (cited, weight 0.31) · coffeetalk.example (formative, weight 0.18)
Not just what AI says. What makes it say it.
Tonecast is a SaaS platform that shows what AI assistants (ChatGPT, Claude, Gemini and Perplexity) say about a brand, product, person or topic, and which web pages shaped those answers.
Monitoring tools stop at the answer. Tonecast follows the answer back to the pages behind it, so you know what to fix.
AI Answers
Visibility, tone and recommendations per engine, prompt and language, sampled weekly through the engines' official APIs. Every number opens the answers behind it.
What shaped it
The source graph: pages the answers cite, pages the engines typically draw on, and your own pages — each with the reason it is there and its influence weight.
Claim check
What AI states about prices, specs and facts, checked against your fact sheet. Wrong and outdated claims come with the pages that keep repeating them.
Early signals
When the tone of the sources moves before the answers do, you see it first. When the evidence is not there yet, Tonecast says so instead of guessing.
Five minutes to set up. No keywords, no source lists.
You describe the subject; Tonecast finds the sources on its own.
Describe the subject
Your brand, product, person or topic, the competitors, and the facts AI should get right.
Choose the questions
Tonecast proposes the prompts real people ask, in every language you track. Keep, edit or add.
We ask, every week
ChatGPT, Claude, Gemini and Perplexity answer; Tonecast reads each answer, its citations and the pages behind them.
Act on the gaps
Prompts where only competitors appear, wrong claims, negative sources with weight: a list you can work through.
Numbers you can defend in a meeting.
Official APIs only
Engines are queried through their official APIs, with the cost of every answer measured. No scraping of closed platforms.
A methodology card
For every campaign: what was asked, how often, on which engines, and how accurate the models behind the scores are.
Quality measured, not assumed
Tone and relevance are checked against labelled data before release, and every score records the model that produced it.
Read the methodology → Compare with listening tools and AI trackers →
Plans that grow with the questions you track.
Prices in euro, VAT excluded where it applies. No free trial: the demo shows everything first.
- 20 tracked prompts · 3 AI engines
- 1 active campaign
- 300 source pages per campaign
- 6 months of history
- Methodology card, PDF report
- 40 tracked prompts · 4 AI engines
- 3 active campaigns
- 1,000 source pages per campaign
- 12 months of history
- Alerts, early warnings, gap & action list, Launch Mode
- 80 tracked prompts · 5 AI engines
- 10 active campaigns
- 3,000 source pages per campaign
- 24 months of history
- Compare, PowerPoint export, shared views
- 200 tracked prompts · 5 AI engines
- 30 active campaigns · 10 workspaces
- 10,000 source pages per workspace
- 24 months of history
- Workspaces per client, white label, API
Full pricing: every limit, the prompt add-on and pricing questions →
Guides, definitions and use cases.
How AI assistants decide what to say about a brand, and how to measure it. Useful whether or not you use Tonecast.
- What is Generative Engine Optimization?GEO explained, and what a brand can do first.
- How to check what ChatGPT says about your brandA step-by-step method you can start today.
- Why AI assistants cite some sourcesRetrieval, citations and the pages behind them.
- Fixing wrong information in AI answersFrom the wrong claim to the page that carries it.
- Measuring AI visibilitySampling, confidence intervals and pitfalls.
- GlossaryAI visibility rate, source graph, FSI, claim check and more.
Good to know
Which AI assistants do you track?
ChatGPT, Claude, Gemini and Perplexity, through their official APIs; Google AI Overviews is on the way. Your plan sets how many engines each campaign uses.
Do I have to tell Tonecast which websites to watch?
No. The sources are found automatically from what the engines cite and from where they typically draw information. You only declare your own sites.
Why is there no free trial?
Every answer we sample has a real cost. The public demo shows every screen with fictitious brands, so you can judge before you buy.
Can an agency manage several clients?
Yes: the Agency plan gives one workspace per client under one company account, with its own members and roles.
What happens if I go over my plan?
Nothing is deleted. New campaigns wait until there is room, and you can upgrade or add prompt packs at any time.
More answers on the pricing, features and how it works pages.