Choose your TruthAI plan

Yes. Truthyx treats information as a public good — access to verified TruthNews articles, TruthFake fact-checks, TruthFeed, TruthMarket, as well as the digital-security training and Microsoft certifications, stays free forever, no subscription. Paid plans unlock only the AI capabilities (advanced chat, personal article generation, images) and pro volumes.

The Pro plan at 7 CHF is sized for a student or a freelancer — cheaper than a news subscription, and it adds the Mistral AI assistant (TruthSmall + TruthMedium models) and 10 verified articles per month. The Premium plan at 15 CHF matches ChatGPT Plus (≈ €20) but adds: the creator volume (50 articles, 150 images) and the TruthLarge model (2.5 M tokens/month).

You choose your model, and the chosen model doesn't change: TruthAI imposes no automatic routing. TruthAI only adds the algorithm that makes the answers to your requests reliable. By default: Mistral, the sovereign European AI (TruthSmall, TruthMedium, TruthLarge and the Magistral reasoning model). On demand, with prepaid credits: Anthropic's Claude — Opus 4.8 with 1M context, Sonnet 4.6, Haiku 4.5 — and OpenAI's ChatGPT (GPT-5.x). Our system-security layer is reinforced by Claude Fable 5, Anthropic's most capable model, coming soon to chat. In short: you access all of these models in a single app, at the TruthAI price — and you decide which one answers.

A general-purpose AI answers confidently without checking, can reflect the biases of its training data, and offers no safeguards for minors. TruthAI adds an algorithm to users' requests so the AI returns a reliable, sourced and safe answer (cross-checked with IFCN fact-checkers), and runs by default on a sovereign European AI (GDPR, hosted in Switzerland + EU), including for its own operation. Our algorithm never trains the models on your prompts. TruthAI also gives you on-demand access to the world's best models — Claude (Anthropic) and ChatGPT (OpenAI) — billed by usage with prepaid credits. Our platform is reinforced by Claude Fable 5, Anthropic's most capable model. For less than a classic AI subscription (~€20), you also pay for reliability and safety.

A classic subscription (~€20) is a flat fee: you pay the same whether you use it a little or a lot, and the provider keeps the difference. TruthAI bills real usage — most members consume far less than a flat plan. Our algorithm's processing adds no extra cost and no extra token consumption on your side. The result: added reliability and safety.

That's our core commitment. The Truthyx algorithm enforces a neutral, non-partisan tone, and every sensitive answer is cross-checked against verified sources rather than an opinion. We don't live off advertising or data reselling: our only revenue is subscriptions — so our only interest is that you get accurate answers.

We want to keep prices flexible — so we can lower them — and readable: a single number per month, no conversion table, no commitment. You can cancel anytime.

Truthyx shows a message offering to either wait for renewal the following month, buy an à-la-carte pack (from 4 CHF), or upgrade to the higher plan. No surprise billing.

No. No prompt is used to train the provider models (Mistral, OpenAI, Claude). Your conversations stay private and encrypted, and don't leave the Truthyx ecosystem: it's TruthAI that talks to the AIs, like a proxy — you only ever talk to TruthAI. Hosted in Switzerland + the European Union, GDPR-compliant.

Yes. Every Truthyx module is available as a native iOS app (TestFlight then App Store, late 2026) and soon Android (Play Store). A single account syncs across all your devices. No need to pay a second time to use the mobile version.