Introduce option to use Google Gemini assistants instead of OpenAI assistants

1. Add the ability to fingerprint the current user in place of a user identifier
2. Add the threads and messages to the database
3. Recreate the auditable chat screen for the original creator of the assistant

Remove the Bring Your Own Key (BYOK) in assistant creation flow. Today it provides an option to bring your own keys for the assistant that should be removed

There is no capture of this information anywhere in the system. Capture only basic information and limited to email address for future communication

Currently only Vercel Blob store is supported. There is desire and intent for us to support Azure, S3, Google Cloud Storage and other blob stores.

This is also required to help support containerization using docker.

Description

To enhance user experience and allow for greater personalization, we propose introducing customization options for the assistant interface. These enhancements aim to provide users with the ability to tailor the assistant's appearance and behavior to better align with their preferences and branding requirements.

Customization Options

The following customization features are proposed:

1. Assistant Image: Allow users to choose a custom image for the assistant avatar, enhancing brand alignment and personalization.
2. Size of the Floating Icon: Provide the ability to adjust the size of the assistant's floating icon, accommodating various screen sizes and user preferences.
3. Open by Default: Offer an option to have the assistant interface open by default when a page loads, ensuring immediate visibility and access.
4. Primary Color: Enable users to define a primary color for the assistant interface, allowing for consistency with brand color schemes.
5. Secondary Color: Similar to the primary color, this option would allow for the customization of a secondary color, further enhancing brand integration.
6. Initial Prompt Message: Allow users to set a custom initial prompt message that the assistant will display, facilitating a more personalized user interaction.
7. Label on the Textfield: Provide the ability to customize the label displayed on the assistant's text input field, enhancing clarity and user guidance.

Objective

The goal of these enhancements is to provide a more immersive and brand-consistent experience for users interacting with the assistant. By offering these customization options, we can significantly improve user engagement and satisfaction.

Implementation Considerations

• Ensure that customization options are easily accessible within the user settings or configuration panel.
• Maintain performance and responsiveness of the assistant interface despite these customization changes.
• Provide default values for all customization options to ensure a seamless experience for users who may not opt for customization.

Expected Impact

• Increased user engagement and satisfaction due to a more personalized and brand-consistent assistant interface.
• Enhanced appeal of the assistant feature to a broader range of users and businesses seeking brand alignment.

We look forward to discussing these proposed enhancements and moving forward with their implementation to improve the user experience.

If the user is on mobile and taps on the assistant, go full screen so there is no clunky text field experience

Initial commit with npx create-next-app@latest
Modify to run with turbo

Ability for users to add and manage files to the assistants

We get an error sometimes in streaming text from OpenAI. This was gpt-4 turbo

https://authjs.dev/getting-started/authentication/email

A lot of free opensource LLM models (listed below) are available in groq_cloud and is super fast, it is best to integrate them

• LLaMA3 8b
• LLaMA3 70b
• Mixtral 8x7b
• Gemma 7b

Need to rethink this functionality and UX. What would be a better way for us to allow users to create new threads without having to force prompting them

Overview

This issue outlines the need for a comprehensive overhaul of our current billing system, aiming to introduce a more transparent and user-friendly billing experience. Key enhancements include the incorporation of a per-token pricing model within our database architecture, seamless integration of payment methods via Stripe during Assistant setup, and the introduction of a dedicated Pricing and Billing page for each Assistant.

Objectives

• Model Table with Pricing Functionality: To provide clarity and transparency in pricing, introduce per-token pricing functionality within the Model table. This will allow users to understand the cost associated with their usage directly.

• Tighter Model-Assistant Association: Ensure a robust linkage between each Model and its corresponding Assistant to streamline functionality and billing accuracy.

• Simplification of Setup Process: Remove the Bring Your Own Key (BYOK) option during the Assistant setup to reduce complexity and potential user errors.

• Stripe Integration for Payment Processing: During Assistant creation, incorporate Stripe to accept and manage payment methods securely and efficiently.

• Dedicated Pricing and Billing Page: Each Assistant should have a tailored Pricing and Billing page, offering users a clear view of their current plan, usage, and billing history.

• Functionality and Feature Management: Temporarily disable any functions or tools that are currently not operational to ensure a clean and user-friendly interface.

• Token Usage Metering: Implement a metering system to track token usage, aiding in billing accuracy and providing valuable insights for users to manage their costs.

Implementation Steps

1. Model Table Enhancement: Update the database schema to include pricing details for each token within the Model table. Ensure compatibility and easy retrieval of this information.

2. Assistant Linkage: Develop backend functionality to tightly couple each Assistant with its corresponding Model, ensuring accurate tracking and billing.

3. Setup Process Revision: Modify the Assistant setup workflow to remove the BYOK feature and integrate Stripe for payment processing.

4. Pricing and Billing Interface: Design and implement a user-friendly Pricing and Billing page accessible from each Assistant's dashboard, displaying essential billing information and controls.

5. Feature Management: Review the current platform features and tools, disabling any that are not yet ready for public use to maintain a streamlined user experience.

6. Usage Tracking: Establish a robust token usage tracking system, ensuring it's accurately reflected in the user's billing and visible on the Pricing and Billing page.

Expected Outcomes

• A more transparent, straightforward billing process that enhances user trust and satisfaction.
• Improved operational efficiency in managing billing and payments.
• Enhanced user experience with the introduction of clear pricing structures and easy payment processing.

Overview

Introduce a feature allowing users to configure the persistence time span of threads. This enhancement will enable users to set the duration for which threads remain active or stored, based on their preferences or requirements, with options ranging from minutes to days.

Feature Description

Currently, our assistants maintains threads for a default duration of 24 hours, which may not align with the varied needs of our users. By providing the ability to adjust the persistence time span of threads, we can offer a more tailored and flexible user experience. This feature would be particularly useful in scenarios where the relevance of information within threads diminishes over time or needs to be retained for longer periods for audit or reference purposes.

User Stories

• As a user, I want to set the thread persistence time to 30 minutes for rapid discussions that quickly become irrelevant.
• As a moderator, I wish to configure certain threads to persist for several days or indefinitely for ongoing discussions or important announcements.
• As an admin, I need the ability to set default persistence times for different categories of threads to better manage content and server resources.

Proposed Solution

• Introduce settings in the user interface to configure the persistence time span of threads, with predefined options (e.g., 30 minutes, 1 hour, 24 hours, 7 days, Indefinitely).
• Allow administrators to set default persistence durations for threads across the platform, with the option for users or moderators to adjust individual threads within those parameters.
• Implement backend logic to automatically archive or delete threads based on their configured persistence settings.

Acceptance Criteria

1. Administrators have the ability to define default persistence settings for all threads.
2. Thread persistence must be scoped to the assistants

Implementation Considerations

• Ensure that the user interface for configuring the persistence time span is intuitive and accessible within the thread creation or settings panel.

Impact

This feature will significantly enhance user and moderator control over content management and potentially improve the application's performance by ensuring that only relevant threads occupy server resources.

We are seeing this warning in both dev and production, causing inefficiencies in production. Will be a problem at scale

Implement the resolution outlined here prisma/prisma#4399

Login with Apple account should be possible - Integrate using NextAuth.js - https://next-auth.js.org/providers/apple

Minimize button should reuse the thread. And close button should warn for prompt to dispose message

Current layout for the chat window is not responsive, it is better to go full screen

So we can asses which conversations are valuable

We need to have metrics for Threads and Messages per assistant

Hello, can you give me an example of the .env that must be deployed and the step by step of the local deploy, so I can see if I can dockerize the project. Can anything else be used with OpenAI and what would the procedure be? @santthosh

Expected results:

Application will show the chat history instead of erroring out

This should be helpful for PRs and allowing contributions

Ability to login with Microsoft using NextAuth.js - More details on this thread nextauthjs/next-auth#1326

Add support for Function Calling in OpenAI assistants as outlined here - https://platform.openai.com/docs/assistants/tools/function-calling

Setup a Vercel pipeline for deployment onto a Demo Application

By default both OpenAI and Google model providers are to be configured. Need to find a way to selectively enable them

https://platform.openai.com/docs/assistants/overview?context=with-streaming

The first few messages are in the right order, but after sometime messages get in the wrong order

Enhance OpenAI agents with File Search capabilities as outlined in the assistants API - https://platform.openai.com/docs/assistants/tools/file-search

Users should be able to create assistants for the organization and be able to transfer assistants between organizations

Stream formats between OpenAI, Google and Groq are much different when sent to the client. There is a potential race condition in when the streamed message is saved and new messages are retrieved.

We need to normalize this and sometimes when results error out, we don't have a great way to handle this.

Description

We need to containerize our Next.js application, which uses a PostgreSQL database, to improve deployment processes and environment consistency. The application is currently hosted on Vercel.

Requirements

1. Dockerization:

   • Develop a Dockerfile for the Next.js application.
   • Use an official PostgreSQL image for the database.
   • Ensure containers can communicate effectively.

2. Environment Configuration:

   • Set up necessary environment variables for development and deployment.
   • Manage secrets securely.

3. Local Development:

   • Provide a docker-compose.yml to facilitate local development with Docker.

4. Deployment:

   • Draft guidelines for deploying the containers to a cloud platform.
   • Suggest a simple CI/CD approach for building and deploying Docker images.

5. Documentation:

   • Update the project documentation to include Docker setup and usage instructions.

Objective

• Enable scalable and flexible deployments using Docker.
• Simplify the setup process for new developers.

Please provide feedback or additional considerations for this initiative.