This guide concerns AI Services, Infomaniak's independent AI platform that allows you to leverage the power of the best open-source AI models on the market without compromising on security and data confidentiality.

Preamble

• Designed for developers and the general public, Infomaniak's cloud provider's AI services ensure data protection through end-to-end processing in Switzerland.
• This platform continuously offers the most powerful open-source AI models on the market (LLM, voice transcription, image generation, etc.).
  • Refer to this other guide to learn more about the lifecycle and migration of models.
• Billed based on usage, this service offers highly competitive pricing.
  • One million free credits allow you to test the service without commitment for a month.

Obtain AI Services

On-demand AI services can be used in three ways:

1. Via the Web app: the operation is similar to the Chat.com, Claude.ai interface, etc. Simply open AI Services and choose the desired AI to address your requests. Infomaniak's chatbot allows you to create custom conversational agents, transcribe audio files, and generate high-quality images in your preferred format (landscape, portrait, square).
2. Via the API: developers and companies can integrate their choice of AI models into their applications and work environments.
   • It is necessary to create an API token.
   • A single product can exist per Organization, but API tokens can be multiple.
   • The API documentation is available at https://developer.infomaniak.com/docs/api and contains information on available methods, parameters, and examples.
3. Infomaniak's AI is also integrated into its tools, notably Site Creator and my kSuite, its collaborative suite that brings together all its online productivity tools:
   • The instant messaging kChat includes a sovereign alternative to ChatGPT for handling daily tasks.
   • kDrive integrates AI, notably for data categorization.
   • The Web app Mail allows you to write, respond to, and improve your emails in one click, while ensuring total data confidentiality.

AI Services can be used by the following users:

• Legal representative of the Organization
• Organization Administrators
• Organization Collaborators with technical rights

To access AI Services and, for example, view your daily consumption:

1. Click here to access the management of your product on the Infomaniak Manager (need help?):
   

Billing

• The API is billed by credit.
• Each request consumes one LLM credit or token *.
• Billing is accessible to legal representatives and administrators.
• Collaborators with technical rights have access to the documentation, but not to the invoices.
• A credit card is required to start using the API.
• An unpaid invoice results in the suspension of the service.
• A spending limit can be set by a user with billing rights in the organization and on the product to limit unexpected costs.
• The rates are presented on the sales page.

* LLM tokens (not to be confused with API tokens) are the basic units used by language models to process and generate text. Each token generally represents a word, a part of a word, or an entity in a text corpus and is used for learning and generating natural language.

This guide explains how to create and manage your authentication tokens to interact securely with the Infomaniak API.

Preamble

• Token features:
  • Number: no limit on creation
  • Validity: unlimited by default (customizable upon creation)
  • Inactivity: a token is automatically deactivated if it is not used for one year
• Usage example:
  • Token for the Newsletter API

Manage API tokens

When creating, you must define the application and the access scope:

1. Access token management on the Manager (need help?).
2. Click on Create a token:
   
3. Select the product/application and the validity period:
   
4. Confirm the operation with your account password.
5. Important: copy the token and keep it in a safe place; it will no longer be displayed afterwards.

To delete, select them and click on the revocation button that appears:

This guide details the creation of illustrations within Ai Services Infomaniak.

Preamble

• Possibility to create images (from 1 to 4 images per generation) from a prompt.
• Possibility to modify certain image parameters such as the format (square, portrait, landscape) and the rendering (more or less details).
• The homepage presents example images that you can draw inspiration from.
• The generated images are available for 30 days in the history.
• The public API of AI Services is used with the models SDXL / flux.

Generate an image with Ai Services

To access the artificial image generation tool:

1. Click here to access the management of your product on the Infomaniak Manager (need help?):
   
    
2. Click on the SDXL language model:
   
3. Write the prompt necessary for modeling your idea and then click on the icon:
   

This guide explains how to customize the type of language used within discussions and exchanges with the Infomaniak AI Services conversational agent.

Preamble

• By default, the responses and elements provided within a conversation with a bot in AI Services will be of the “usual” type, the assistant is named “Infomaniak Assistant” and its description is clear:
  
• It is entirely possible to invent and customize to the extreme a way that must be followed to the letter by the bot during exchanges.
• Once the method is saved, you can reuse it and switch from one to the other as needed.

Customize the AI Services language model

To create a new “system prompt” and use it:

1. Click here to access the management of your product on the Infomaniak Manager (need help?):
   
2. Click on a language model to start configuring a conversation:
   
3. Click on Modify the current conversational assistant.
4. Click on ‍ New below the choice of the conversational assistant:
   
5. Give a title to the oriented conversation you are configuring.
6. Write the prompt necessary for configuring your bot, detailing whether it should behave like a terminal, a specialized translator, etc. (examples of prompt)
7. Save the changes:
   

You can now start a conversation and get different responses or elements depending on the behavior inherent in the chosen system prompt:

It is also possible to choose a model for each conversation (refer to this other guide if you are looking for information on models and responsibilities).

All these new parameters are available in a side panel accessible on the right side of the screen from all conversations.