Last updated: 15 November 2026
This AI Disclosure explains, in plain language, how artificial intelligence is used in journail.app and the Journail mobile app (the "Service"). It is provided by Tomaž Pernovšek s.p. ("we", "us").
It is our transparency notice under Regulation (EU) 2024/1689 (the EU AI Act). The transparency obligations of Article 50 apply from 2 August 2026, and Journail meets them as described below.
1. You are talking to an AI
When you use the AI Assistant ("AI pomočnik"), the guided morning planning conversation, or the evening journal close, you are interacting with an AI system, not a human. Nobody at Journail reads or writes your replies.
If you are ever unsure whether something in Journail was written by AI or by a person, assume it was written by AI, and ask us at info@journail.app.
2. Risk classification under the EU AI Act
- Not a prohibited practice (Article 5). Journail does not manipulate behaviour, exploit vulnerabilities, score people socially, recognise emotions, or perform biometric identification or categorisation.
- Not high-risk (Annex III). Journail is not used in employment, education admission, credit, insurance, essential services, law enforcement, migration, justice or any other Annex III area, and it is not a safety component of a regulated product.
- Limited risk — transparency obligations (Article 50). Journail is a conversational AI system for personal planning and journaling, so the obligations that apply are to tell you that you are interacting with AI and to make AI-generated text recognisable.
Our role. Journail is a deployer that integrates general-purpose AI models made by third-party providers. We do not train our own foundation model. Current default models come from providers such as OpenAI (GPT family) and Anthropic (Claude), and we may change models as better or safer ones become available. Models are accessed through OpenRouter, an AI gateway acting as our data processor.
3. What the AI does in Journail
3.1 AI Assistant
A side conversation available across the app. It answers questions about your Journail data only — your plans, tasks, calendar, goals and journal — and politely declines unrelated topics. How far back it can see depends on your plan:
| Plan | Journal context the AI Assistant can use |
|---|---|
| Free | Last 7 days |
| Plus | Last 30 days |
| Premium | Your full history |
3.2 Guided morning planning
A short conversation about your day. The AI reads your active goals, today's calendar and your tasks, and proposes a focused plan. Proposed changes to tasks or calendar are shown to you first and applied only when you confirm.
3.3 Evening journal close
The AI asks a few questions about how the day went and then drafts a journal entry from that conversation. The draft is shown to you as an AI draft: you can edit it, accept it, or discard it. Nothing is saved to your journal until you confirm.
3.4 Weekly and monthly reviews (Premium)
Periodic AI reviews that summarise your entries and point out patterns (for example recurring themes or goals that keep slipping). They appear in the app and, if you allow it, by email and push notification.
3.5 Morning and evening email briefs (Plus and Premium)
Short AI-written emails that summarise your day ahead or the day behind. You can turn them off in your email settings.
3.6 Task and goal suggestions
The AI may suggest new tasks, rescheduling, or links between tasks and goals. These are suggestions only.
4. How we make AI recognisable (Article 50)
- Notice in conversations. The top of the AI Assistant and of the planning and journal conversations states that replies are generated by AI and may be wrong.
- "Written by AI" labels. Weekly and monthly reviews and AI-written emails (briefs and reviews) are labelled as written by AI.
- AI drafts. A journal entry drafted by the AI is presented as an AI draft that you review, edit and confirm. Once you confirm it, it is your entry; Journail does not keep a permanent per-entry "AI" tag.
- No synthetic media. Journail does not generate images, audio or video, so deepfake disclosure and watermarking obligations (Article 50(2) and (4) for synthetic media) do not arise.
- Nothing is published. AI-generated text stays in your private account. Journail does not publish it to the public, so the obligation to label AI text published to inform the public does not apply.
- This page. This disclosure is linked from the site footer and the legal pages.
5. Human oversight — you stay in control
- Nothing changes without your confirmation. The AI cannot create, edit, move or delete tasks, calendar events, goals or journal entries on its own. Every change it proposes is shown to you and applied only after you confirm.
- No automated decisions about you. The AI makes no decision with legal or similarly significant effects within the meaning of Article 22 GDPR.
- You can work without AI. The planner, task manager, calendar, goals and manual journal work without any AI feature.
6. Your consent to AI processing
Before you use any AI feature for the first time, Journail asks for your explicit consent to send the relevant content to our AI providers. Without that consent, AI features stay off.
You can withdraw consent at any time in Settings. AI features are then switched off; withdrawing does not affect processing that already happened.
7. Your data and the AI providers
- Never used for training. Every request we send is routed with OpenRouter's
data_collection: "deny"setting, so it only reaches model providers that do not collect or train on API content. Your content is never used to train, fine-tune or evaluate AI models. - Only what is needed. Each request contains your message, the slice of context the feature needs (within your plan's limits above), and instructions that define the AI's behaviour. We do not send your email address or payment details.
- Transient retention. Providers may keep requests briefly for abuse monitoring and operational stability (typically up to 30 days) before deleting them.
- Encryption. Data is encrypted in transit and at rest.
Our full list of sub-processors is in the Privacy Policy.
8. Safety
Journail is not a therapy, medical or crisis service. Its AI is not a professional and its output is not medical, psychological, legal or financial advice.
Journail runs a narrow crisis-language detector (Slovenian and English keywords). When a message contains language that may indicate a crisis, the AI receives an additional safety instruction to respond with care, encourage you to reach out to people who can help, and share emergency and helpline numbers. The detector is a simple keyword check: it can miss things and it can trigger by mistake. It does not alert anyone, and no one at Journail is notified.
If you or someone else is in danger, call 112 (EU) or your local emergency number. See our Safety page for helplines.
9. Limitations
- The AI can be wrong. It can misread your context, invent details, or give poor suggestions. Check anything important.
- It reflects what you share. Its suggestions are only as good as the goals, tasks and entries it can see.
- It is not neutral about everything. Language models can carry biases from their training data. Tell us if a reply seems unfair or inappropriate.
10. AI literacy, quality and feedback
- AI literacy (Article 4). The operator of Journail keeps up to date with how the models used work, their limits and risks, and the obligations of the EU AI Act.
- Quality monitoring without reading your content. We monitor AI quality and cost through technical metrics only — token counts, error rates, response times and feature usage. We do not read your conversations or journal entries to do this.
- Feedback and complaints. If an AI reply was wrong, harmful, distressing or off-topic, write to info@journail.app. You can also lodge a complaint with the Slovenian Information Commissioner (Informacijski pooblaščenec, www.ip-rs.si) or your local supervisory authority.
11. Changes to this disclosure
When we change AI providers in a material way, change what data is sent, or add a new AI feature, we update this page and the Privacy Policy, and notify you in the app or by email where the change is material.
12. Contact
Tomaž Pernovšek s.p. Dolinškova ulica 10a, 1000 Ljubljana, Slovenia Email: info@journail.app