A gem with exciting potential
I’ve found remio AI genuinely helpful, and it has made a real difference in how I work. I especially appreciate having my information and AI assistance together in one place.
One capability that would make remio even more powerful is configurable, multi-step workflows with specialized subagents coordinated by a main agent. For example, in a software sprint, a planning agent could turn requirements into tickets, developer agents could implement them and run unit tests, and a testing agent could check each result against the original requirements. If something fails, the tester could send it back to development with specific issues to fix. Users could set a limit on how many development-and-testing rounds happen before the workflow asks for their input.
I currently try to approximate this by keeping a project folder with a set of README.md files that define the agents and how they should work together, but this workaround is cumbersome and doesn’t work very well. The coordinator could manage assignments and handoffs, make sure each agent has the right context, track progress, and escalate decisions that need the user’s approval.
It would also be useful for each agent to retain context about its previous work within a project, so it can pick up where it left off instead of unnecessarily repeating checks or tasks. A shared project record—with the requirements, tickets, decisions, test results, and handoff history—could remain the source of truth.
Some of this may already be possible through workarounds or may be on the team’s roadmap. I’m sharing this as a use case because I think persistent, user-configurable workflows could make remio even more useful for repeatable, end-to-end work. Thanks for building and continuing to improve remio.
As a separate, smaller feature request, it would be useful to add tags to chat groups. That would make it easier to organize and find relevant past conversations, so I could refer back to them instead of searching online for the same information again.