User types and long term memory

[vontainment]

Self Checks

• [X] I have searched for existing issues search for existing issues, including closed ones.
• [X] I confirm that I am using English to file this report (我已阅读并同意 Language Policy).

1. Is this request related to a challenge you're experiencing?

No

2. Describe the feature you'd like to see

You think the regular self hosted version should be able to have multi workspaces using email password auth.

Then at an option to bots for long-term memory. It can either be attached to the workspace or attached to users. Basically have it store the conversations in the vector database.

Should be an option to enable/disable public sign up. Should be an option to limit it to particular domain names for email or not. Then people can sign up with their email and a password. And it would save their conversations privately.

3. How will this feature improve your workflow or experience?

It allow for a more personalized bots

4. Additional context or comments

No response

5. Can you help us with this feature?

• [ ] I am interested in contributing to this feature.

[dosubot[bot]]

🤖

Hey @vontainment! Nice to see you back. Hope all is well with you. Let's dive into this interesting idea you've brought to the table.

Thank you for your feature request. Based on the current state of the dify repository, here's what I can tell:

1. Multi-workspace functionality: The current version of dify does support multi-workspace functionality. This is evident from the WorkspacesContext in the workspace-context.tsx file, which is designed to hold an array of workspaces and fetch them from the '/workspaces' endpoint.

2. Email password authentication for multi-workspaces: Currently, dify does not support email password authentication for multi-workspaces. The AccountPasswordApi class in the account.py file allows for updating the password of the current user, but it does not specify anything about multi-workspace support.

3. Storage of conversation history in the vector database for long-term memory: The current version of dify does support the storage of conversation history. This is evident from the pagination_by_first_id and pagination_by_last_id methods in the MessageService class in the message_service.py file, which query the database for messages associated with a specific conversation. However, it's important to note that this doesn't necessarily mean the conversation history is stored in a vector database for long-term memory. The code provided doesn't specify the type of database being used, only that the conversation history is stored and can be retrieved.

Your feature request for email password authentication for multi-workspaces and the storage of conversation history in the vector database for long-term memory is valid and would indeed enhance the functionality of dify. However, as a Dosu, I don't have the ability to create pull requests or issues. I recommend keeping an eye on the repository for updates or changes that might address these features.

Sources

• web/context/workspace-context.tsx
• api/controllers/console/workspace/account.py
• api/services/message_service.py

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[crazywoola]

You think the regular self hosted version should be able to have multi workspaces using email password auth.

For self-hosted version, multi-tenancy requires our business license, you could send an email to [email protected] to get an quota.

[github-actions[bot]]

Close due to it's no longer active, if you have any questions, you can reopen it.

[takatost]

long term memory will be a part of our future roadmap , and currently we are in the process of building up that capability.

[github-actions[bot]]

Close due to it's no longer active, if you have any questions, you can reopen it.