Reviewers mostly see Boost.space as a strong data hub and automation layer that pulls scattered app data into one reliable system, cutting manual work and keeping teams in sync. Users repeatedly praise its flexibility, broad integrations, Make-based workflows, and helpful support, with some calling setup surprisingly fast. The main drawback is a real learning curve: several reviewers say the platform can feel overwhelming at first and needs better beginner tutorials, deeper documentation, and in a few cases more local integrations or more accessible support hours.
We run a lot of agents across different tools, and the hardest part for us is knowing which agent is working on what, and why. Does AgentOS keep a shared "goal" or context that every agent reads before it starts, or is that up to each agent's setup?
@tsubasatakeuhi Thats actually on of the core foundations of AgentOS. Since you got all agents running via one platform you can see IF they are working and ON WHAT and use Boosty - the agent coordinator to keep everything aligned and you informed.
@matous_kralik thanks! seeing who's on what from one place is exactly what we were missing. will try Boosty
@tsubasatakeuhi Let me know how it went! 🙏🔥
@matous_kralik will do, thanks! 🙌
Starting with one job feels like a sensible way into this. I'd want to see which shared instructions an agent used for a result, especially after feedback has changed its skills.
@oleksii_sekundant Thanks for jumping in. You can see each agents skills, what skills it used for the task and then evaluate the skill itself and its ROI. Agents do change the skills on their own but to be honest I always make sure and double down on the change myself - meaning telling the agent "Hey, based on X change the skill Y" and wait for the confirmation.
How will approval points behave for scheduled runs when nobody is online? Does the agent pause only the step waiting for sign-off and carry on with the rest of the task, or does the whole task stop until someone approves it in Slack?