AI Transparency Notice
Effective and last updated: 22 October 2026
Intelligence, live voice, insights, period reports, market reports and the support assistant are AI systems. This notice explains what that means for you, which companies process your prompts, what is synthetic and what you should not rely on. It is published under the transparency rules of the EU Artificial Intelligence Act and comparable laws, and it forms part of the Terms of Service and Privacy Policy.
1. You are interacting with an AI#
Every reply in Intelligence, every spoken answer in live voice mode, every automatically generated insight, period report and Market Monitor report, and the first-line support bot are produced by large language models, not by a person. Where a person joins a conversation (for example a support agent taking over a chat, or an administrator reviewing an appeal) the Service says so.
Synthetic speech: the voice you hear in live voice mode is generated speech. It does not belong to a real person, is not a recording and is not the voice of an adviser. Voice characters are labelled as synthetic in the voice picker.
Generated images and text: share cards, campaign images and AI-written summaries are machine-generated from your data. If you publish them, you are responsible for how they are presented.
2. What the AI is for, and what it is not#
The AI is an educational coaching and analysis tool that works from your own journal. It is not an investment adviser, does not know your full financial situation, cannot see live order books and does not execute trades. Its output can be wrong, incomplete, out of date or confidently misleading, especially about numbers, prices, dates and symbols. Verify anything you intend to act on against a primary source. Nothing it says is financial, investment, tax or legal advice.
No automated decisions about you, except safety
Intelligence and the other coaching features never make decisions that produce legal or similarly significant effects about you. The only automated decisions on the Service are safety enforcement actions (warnings, restrictions, suspensions) described in the Acceptable Use and Community Guidelines; they are applied only when evidence and confidence thresholds are met, and a person reviews every appeal.3. Who processes your prompts#
Prompts, attached images, relevant conversation history and the journal context the feature selects are sent to one of the following providers to generate a reply. Which one depends on your plan, the task and availability; the model and provider are shown in the conversation metadata. Each provider processes the content only to return a response for you; none of them is permitted to use it to train their models under our agreements.
| Provider | Used for | Location |
|---|---|---|
| OpenAI | Premium model for chat, insights, reports, vision and voice turns | United States |
| Mistral AI | Backup model for chat, insights, reports, safety reviews and memory search when the other providers are busy or down; speech-to-text and text-to-speech for live voice; safety classification | France (EU) |
| OpenRouter | Routing to free-lane models for chat, insights and reports | United States |
| NVIDIA | Free-lane model inference when OpenRouter is unavailable | United States |
| Tavily | Web search when an answer needs current public information; only a search query derived from your request is sent | United States |
The platform never pays for a user's AI usage beyond the plan: Premium usage draws on the monthly credits, then purchased credits, then free models. The free lane is available to everyone.
4. What is stored and how to control it#
- Conversations are stored in your account so you can return to them. Delete any chat or project at any time.
- AI request records keep metadata about each model call (the feature, provider and model, token counts, timing and the result), never the prompt or the answer. They are kept for 90 days to run the AI lanes, reconcile credit and spot abuse, and are deleted with your account. See the Privacy Policy.
- Memory (lasting facts about your trading) is generated from chats only if the switch in Settings → Intelligence → Memory is on. You can read, edit and delete every memory, turn off generation, stop memory being used in conversations, and decide whether sensitive topics (health, religion, politics, identity) may ever be remembered. That last switch is off unless you turn it on.
- Voice audio is processed in real time to produce a transcript and reply and is not kept as a recording.
- Learning examples: we may create de-identified examples from selected interactions to evaluate quality and ground later answers. Identifiers, amounts, dates and links are removed or generalised first. Deleting your account removes your conversations; de-identified examples that no longer identify you may be kept.
- Feedback you give on a reply (thumbs up or down, notes) is stored with the reply to improve routing and quality.
5. Known limitations#
- Models may hallucinate facts, misread screenshots, confuse symbols or time zones and mis-hear numbers in voice mode.
- Market reports are generated per instrument and may be shared across users; they are not tailored to your position.
- Web research retrieves bounded excerpts from public pages and cites them; sources can be wrong or outdated.
- Quality differs between the premium model and free-lane models; limits and quotas apply and are shown in Settings → Intelligence → Usage.
- Live voice depends on network conditions and provider capacity and may be interrupted.
6. Prohibited uses of the AI#
You may not use the AI features to make decisions with legal or similarly significant effects about another person, to give regulated advice to others without the required licence and human review, to clone a voice or likeness, to generate illegal, infringing, deceptive or harmful content, or to extract model weights, prompts or other protected components. See the Acceptable Use and Community Guidelines.
Who we are and how to contact us
TheProfitPath
The operator's registered name and postal address are being finalised and will be published here. Until then every notice below reaches the operator directly.
- Legal notices and disputes: [email protected]
- Privacy requests: [email protected] or the privacy request form
- Illegal content, abuse and copyright notices: [email protected] or the report content form
- Digital Services Act point of contact for authorities and users: [email protected] (English, Slovenian)
- Security vulnerabilities: [email protected]
- Everything else: contact support or the in-app support chat
All legal documents
- Terms of ServiceThe agreement that governs your account, subscriptions, AI features, community and the developer API.
- Privacy PolicyWhat personal information we collect, why, who receives it, how long we keep it and the rights you have wherever you live.
- Cookie PolicyEvery cookie and browser-storage key the Service sets, what it does and how to control it.
- Refunds, Cancellation and WithdrawalHow to cancel, when a refund applies, the 14-day withdrawal right for EU and UK consumers and the model withdrawal form.
- Acceptable Use and Community GuidelinesWhat you may not do on the Service, how content is moderated, how to report content and how to appeal a decision.
- AI Transparency NoticeThis page
- Data Processing AddendumThe processor terms that apply when an organization uses the Service with its members' personal data.
- SubprocessorsThe third parties that process personal information on our behalf, where they are and what they do.
- Accessibility StatementOur accessibility target, known limitations and how to report a barrier.
- Legal NoticeWho operates TheProfitPath, how to reach us for legal, privacy, copyright and safety matters, and the full list of legal documents.
- Report contentReport illegal content, abuse, impersonation or a copyright problem.
- Privacy requestExercise a data protection right without signing in.
- Cancel subscriptionCancel a paid plan without signing in.