How to create a new app?
Last updated
Last updated
Begin by accessing your platform's dashboard. Locate the βAppsβ page from the main menu to manage existing apps or create new ones. To start the app creation process, click the βCreateβ button displayed on the top right corner of the βAppsβ page.
Select the type of app that best fits your needs from the available options. Your choice will determine the app's basic structure and capabilities.
Then, determine who will have access to your app:
Public app: The app link will be public, and anyone with the link can use the app. Ideal for applications meant for a broad audience.
Private app: The app link will be private, and only those with the link can access the application. Suitable for internal or restricted use cases.
Choose an orchestration method from the following options:
Knowledge Base (Basic): In this method, the application relies on the provided knowledge base to answer queries. This is a straightforward method using pre-existing information.
Flow (Advanced): This method allows you to design custom flows for handling queries, allowing more complex interactions and processes. This option is ideal for advanced users who need specific logic and sequences.
Next, enter a unique name for your application and provide a description if desired. This name and description help identify the app and can be changed later if needed.
Next, from the available list, choose one or multiple knowledge bases that you want to connect to your app. You can use the search function to find specific knowledge bases from the list.
Users have the option to select one or multiple schemas linked to the chosen knowledge base to conduct detailed data queries.
Define particular instructions or guidelines that the app must adhere to during user interactions. These instructions may encompass the following: Be specific and detailed: Provide clear and comprehensive answers to user queries. Ensure that responses address specific aspects of the question or topic raised. Use natural language: Interact conversationally with users, using language that feels natural and engaging. Avoid overly technical jargon unless necessary for clarity. Tailor responses to context: Adapt answers based on the conversation's context. Use previous interactions to provide more relevant and personalized responses.
Users can modify the settings of the application to customize default configurations. This can be done by accessing the βEditβ button, which allows for manual adjustments of various parameters. These parameters are:
Model: Choose an apt large language model to power your app.
Temp: Adjust the temperature setting from 0.1 to 1 to control the randomness of the model's responses.
Score threshold: Set the score threshold from 0.1 to 1 to match the query context with the knowledge base.
Context length: Define the length of the context the model uses to generate responses.
Context max token: Set the maximum number of tokens allowed in the context.
Response max tokens: Specify the maximum number of tokens the model can use in its response.
Followup conversation: Enable or disable the modelβs ability to remember and refer back to previous conversations within the same session.
Use CSV agent: Enable or disable the use of a CSV agent to handle structured data inputs and outputs.
Summarize docs: Allow the model to summarize documents as part of its responses.
Source: Enabling this allows the app to specify the source for the information.
Chunking strategy: Enable or disable the chunking strategy to break down documents into smaller, manageable chunks for improved processing.
Collect user info: Enable this to prompt the user for their email or other identifying information when they open the app.
Previous session for identified users: When enabled, the app retrieves the latest queries from the database for the identified user and includes those previous queries in the model's context for highly personalized responses.
Static: When enabled, the app will prompt for a static text input, which will be sent to the model for every query as part of the prompt. This can be used to provide consistent context or instructions to the model.
Once changes are made in the manual configuration section, two buttons are available:
Default settings: Click to revert to the default configurations.
Update settings: Click to save the customized configurations for future use.
By clicking βUpdate Settings,β the customized configuration will be saved and applied to the application going forward.
You can customize the visual and interactive aspects of your app by configuring the following:
Welcome message: Add a welcome message that users will see when they first access the app.
App name and description: Provide a name and a brief description of your app.
Sample questions: Add up to nine sample questions to guide users on how to interact with the bot. You can also generate these questions automatically.
Upload logo and app theme: Enhance your appβs branding by uploading a logo and selecting a theme.
Bot name and icons: Customize the bot's name and icons to match your appβs branding.
After clicking βNext,β your app will be created, and you'll be directed to an overview page where you can manage various aspects of your application.