Feature Request

Currently, our structure is a workspace which serves as a collection of embedded spaces. It would be useful to be able to collect these spaces in sub-workspaces. This could be useful in the parent-child relationships described in other feature requests, but would also be useful as a mechanism for organizing work. The current flat hierarchy begins to buckle under heavy usage.

Currently, I can split a space by cloning it. Occasionally, work being done in multiple spaces comes together. It would be useful to be able to merge spaces, combining them, so that subsequent work could draw on what was in each of the prior spaces. This could be handled through a parent-child relationship, but an MVP implementation would simply require the ability to merge spaces or potentially clone a space and then merge the cloned space with another space.

Currently I can only see one workspace at a time. It would be useful to be able to have multiple workspaces visible in the command bar to facilitate moving between them or moving spaces between workspaces.

Most of the current LLMs have a tendency to go off track on deep research. The normal secondary query to any deep research request is simply, "Are you sure?" It would be helpful if this prompt could be automatically added to the researcher agents.

Clicking on the three dots gives you the opportunity to "edit the chat," so the title is somewhat misleading.

Collaborating with an AI provides some interesting differences from collaborating with humans. The biggest difference is probably around the nature of collaboration. Humans typically talk. AIs communicate through text, but that text is generally written rather than spoken. In a collaborative environment, if we wish to collaborate with the AI, we need the ability to take notes and communicate those notes with the AI. Sometimes those notes are shared on a strictly interactive basis, in which case the Medullar UI/UX works very well. But sometimes we wish to take extended notes - thoughts on a conversation, perhaps a series of questions. Recording these into a chat is frequently not the best way to share them either with fellow collaborators or with the A.I. Storing them in a separate note-taking app is problematic. If that app happens to store its output in a place accessible to Medullar, all is good, or at least sort of good. The AI can have access to those notes, but there is no real way to highlight that they are part of the space. What's needed is an easy-to-use and read interactive note-taking capability within the Medullar app. It would be great if that note-taking capability permitted private notes, but the essential feature would be to take notes easily in day-to-day processes and then have those notes intrinsically available to Medullar. The note-taking app wouldn't need a ton of sophistication. That sophistication is provided by the AI and by Medullar. A multi-page or section outlining app would be near perfect. Something like Microsoft OneNote would be way more complex than is actually needed.

Currently, we mix agent selection with LLM selection in a single drop-down list. These two capabilities are intrinsically separate, and we should be able to run a researcher or a sales researcher or whatever against either ChatGPT or Grok or any other LLM that you add to the product. We should also probably seriously think about how we select agents. The ability to understand the capability of the agent and even enhance that agent rather than simply seeing the name in a list will become progressively more important.