Last updated: May 2026
Otti is an AI-powered A-Level tutor built into the LearnWithOtti platform. Otti is trained on the specification, mark schemes and command words for each subject and exam board we support, as listed on our website. Otti is an automated AI system, not a human. It does not have a physical presence and is not a real person named Otti.
Otti is built with Anthropic, a large language model refined by the team at LearnWithOtti. You can find out more about Anthropic at anthropic.com.
A large language model (LLM) generates text by predicting the most likely next words based on patterns learned from a large corpus of training data. It does not think, reason, or understand in the human sense. Responses are generated statistically, not retrieved from a database of verified facts.
Before every conversation, Otti receives a system prompt written by the LearnWithOtti team that instructs it to behave in particular ways. These instructions include:
Like all language models, Otti can produce responses that contain factual errors, oversimplifications, or “hallucinations” — plausible-sounding content that is incorrect. Always cross-reference important information with your textbook, teacher, or official exam board resources.
Otti’s knowledge is based on its training data, which has a fixed cut-off date. It is not connected to the internet and cannot access current news, data, or specification updates in real time.
Otti provides indicative feedback based on mark-scheme principles. This feedback is not official examiner marking and should not be treated as a guarantee of how your answer would be graded in a real exam.
Otti is not a qualified teacher, examiner, or educational institution. LearnWithOtti is not affiliated with, endorsed by, or accredited by Edexcel, AQA, or any other exam board.
Otti does not retain memory of previous conversations. Each new session starts fresh. Your progress data is stored separately by LearnWithOtti and may be used to contextualise new sessions, but Otti does not have access to prior conversation transcripts by default.
Otti cannot search the web, access external websites, or retrieve up-to-date information from external sources.
When you send a message to Otti, the content of that message is transmitted to Anthropic’s API to generate a response. We do not transmit your name, email address, or other personally identifying information to Anthropic as part of this process.
Anthropic processes message content in accordance with their own privacy policy, available at anthropic.com/privacy. We recommend reviewing this policy to understand how your message content is handled by Anthropic.
Your conversation history within a session may be stored in our database to power your progress tracking and to allow you to review past sessions. See our Privacy Policy for full data retention details.
We may use fully anonymised interaction data (data that cannot be linked back to you) to improve Otti’s prompts, identify common learning gaps, and develop educational content.
LearnWithOtti uses automated processing to calculate your progress scores and topic completion data. This automated processing is used solely to provide you with feedback and tracking within the platform. It is not shared with schools, exam boards, universities, or any third party, and does not produce legally significant decisions about you.
If you encounter a response from Otti that you believe is factually incorrect or inappropriate, please report it to us at rob@learnwithotti.co.uk. Include the subject, topic, and a copy of the response if possible. We take all error reports seriously and use them to improve the Service.
We will update this AI Transparency Notice when we make significant changes to how Otti works or how we handle AI-generated content. The “Last updated” date at the top of this page shows when it was last revised.
For any questions about Otti or this AI Transparency Notice, contact us at rob@learnwithotti.co.uk.