Legal notice on the use of the Mealjam AI beta

1. handling of conversation data and technical implementation

1.1 Conversation data

We use stored conversation data to provide you with an optimal and personalized use of the Mealjam AI. This includes in particular

​

• the provision of an ongoing conversation history over several sessions,

• improving the relevance of responses based on previous interactions

• anonymized analysis for the further development of AI functions,

• the recognition of usage patterns in order to continuously optimize our offer.

​

Interaction decisions, topic interests and user behavior are processed exclusively to improve your individual user experience.

​

1.2 Technical implementation

Mealjam AI processes your requests via a language model**(e.g. ChatGPT Nano / GPT-4o mini / GPT-5**), which is provided via the interface of a technical service provider.

The following data can be transmitted:

​

• Your entries (text prompts) in the chat,

• for logged-in users: Your name or user ID for personalized addresses,

• contextual information such as recipes viewed, meal plans or interests,

• with optional use of voice input: temporary audio data for text conversion.

​

1.3 Use of Typesense and access to Mealjam database

To improve search and response quality, Mealjam also uses the Typesense service as a fast search and indexing service.

​

• Typesense has access to selected structured data from our Mealjam database (e.g. recipes, ingredients, tags, meal plans).

• This data is processed in the form of search indices to enable the AI assistant to provide contextually accurate and fast answers to your queries.

• In addition, a specially connected AI model ("ChatGPT Nano") is used, which receives access to relevant Mealjam content via an integrated Typesense interface in order to generate context-based answers.

• No personal user data is transmitted to Typesense unless this is technically necessary or expressly triggered by a user input.

​

This combination of the AI model and Typesense makes it possible to provide you with faster, more precise answers tailored to your usage.

​

1.4 Web services

Mealjam can use optional web services for up-to-date content (e.g. Tavily Search API) to enhance the information. Mealjam does not currently use any external web search service for information research.

​

1.5 External services

Terms of use of the services used:

​

• OpenAI: https://openai.com/de-DE/policies/eu-terms-of-use/

• Typesense: https://cloud.typesense.org/legal/terms

• Azure: https://www.microsoft.com/de-de/rechtliche-hinweise/nutzungsbedingungen

​

2 General terms of use

2.1 Scope of services

The Mealjam AI supports you with questions about nutrition, recipes, meal plans and the use of the platform. For example, it can suggest recipe ideas, adapt meal plans or provide general nutritional information.

​

2.2 Limitations of AI technology

AI systems can provide incorrect or incomplete information ("hallucinations"). Therefore, always check answers for accuracy and appropriateness.

​

2.3 No medical advice

Mealjam AI is not a substitute for medical or nutritional advice.

​

2.4 No legally binding declarations

No legally binding declarations can be made via the AI.

​

2.5 Copyright

Do not enter any copyrighted content in the chat.

​

2.6 Prohibition of discriminatory content

The use of illegal, discriminatory or defamatory content is prohibited.

​

2.7 Human decision-making authority

The AI provides support, but does not make any decisions for you.

​

2.8 Confidentiality and data protection

Avoid personal and confidential information in the chat.

​

2.9 Business and trade secrets

Please do not enter any confidential company information.

​

2.10 Lawful use

Use may only take place within the framework of applicable laws.

​

2.11 Automated use

Automated queries, scraping or data mining are prohibited.

​

3. Liability

KiezCode GmbH is liable within the scope of the statutory provisions in accordance with the general liability regulations in the Terms of use. 

KiezCode assumes no liability for the accuracy, completeness or timeliness of the information generated by the AI.