Introduction

https://www.loom.com/share/30f2b97c0cb4425784f92fd53d2d39da?sid=12fee583-3e09-4873-b37a-3bd56bfd9ea2

Current Limitations

Currently, Unriddle restricts each data source (e.g. an uploaded PDF or document) to one continuous chat thread. All questions and answers about that source are forced into one long conversation. This is limiting because users often have multiple questions for the same document.

For example, a researcher analyzing a lengthy report might want to discuss methodology questions in one thread and summary insights in another. With the single-thread design, these diverse topics get jumbled together, leading to a cluttered and unfocused chat. Important points become hard to track, and context from one question can undesirably bleed into another. In short, the one-thread-per-document model hampers the user’s ability to organize their thoughts when working with a rich or complex data source.

Proposed Feature: Multiple Threads

Allow users to create multiple independent chat threads all linked to the same document or data source. Instead of one monolithic conversation, a user could start a new thread for each separate topic or question series regarding that document. All these threads would live under a shared container (a "Project" or "Discussion Set") tied to the data source, keeping them united in the workspace without mixing their content.

How It Works:

• New Thread Creation. In the UI, when viewing a document’s chat, the user can click a simple “New Thread” button to spawn a fresh, empty conversation (just like starting a new chat, but without re-uploading the file). Each thread functions like an independent chat session with the AI, but all use the same underlying document as context.
• Project. All threads for a given data source are grouped together.
• Cross-Reference Links. Users can link between threads to reference information. For instance, Thread 2 could have a note like “See insights from @Thread 1”. This is a feature for advanced use, allowing conversations to point to each other without merging them. The threads remain separate by default, but the user has a way to connect the dots if needed.

Notably, these changes would be designed to minimize UI upheaval. The interface might simply show a list or tabs for multiple discussions under one document, analogous to how one might navigate multiple open tabs in a browser.

Benefits and Impact on User Experience

Adopting multi-threaded chats for a single data source would unlock several key benefits:

• Improved Knowledge Organization. Users can keep different topics or questions in separate conversations, all under one project. This makes it far easier to navigate and review past discussions.
• Better Context Management. With distinct threads, each conversation’s context remains narrow and relevant.
• Enhanced Collaboration: In team scenarios or shared projects, multiple threads empower parallel discussions. Team members can explore different questions simultaneously (each in their own thread) while all remaining within the same project space.
• Increased User Retention and Engagement: Ultimately, giving users more control over how they structure their inquiries will make Unriddle more indispensable for complex tasks.