Joao Seabra

The founder's insight here is dead on. Every language app nails vocabulary and grammar but skips the thing that actually builds fluency: forcing yourself to speak in real, unpredictable conversations. That gap is where most learners stall. The real scenarios angle is smart too. "Order a coffee" or "meet someone for the first time" gives you a concrete goal to hit, which is way more motivating...

The "package manager for agent skills" framing clicks immediately, especially coming from the Snyk founder. The dependency management and security signal problem in traditional code is exactly what's now happening with agent skills, and most teams don't have the tooling to even see it yet. The ElevenLabs 2x result is a concrete proof point that avoids the usual vague benchmark claims. That kind...

Instruction adherence in real-time voice is the unsexy problem that actually determines whether voice agents ship to production or stay in demos. Good to see this getting serious attention. The multilingual accuracy improvement is the one I'm most curious about. Does it hold up equally across languages or are some still significantly behind English? That gap tends to be what blocks voice AI...

"Think Stripe for AI agents" is a sharp analogy and it lands well. One script tag to make your site agentic is the kind of primitive that could quietly become infrastructure for a lot of products. The checkout completion use case is compelling because it hits a real pain point. Curious how Rover handles sites with heavy dynamic rendering or multi-step auth flows. Those are usually where browser...

The branding is doing a lot of heavy lifting here and it works perfectly. "No nudging. No reminders. Just follow-through." is a great line because it calls out exactly why every other productivity app fails. As a parent of two young kids, I actually find the Mom angle hilarious and relatable in equal measure. The angry cartoon face is funnier than it has any right to be. Genuine question...

The dynamic routing piece is what catches my eye here, letting the agent decide which tool to call based on the goal rather than hardcoding the path is the right architecture for real-world workflows. Most no-code AI builders skip this and it shows. Memory + Interactive Chat in the same loop is a meaningful combo. Does the memory layer persist across separate workflow runs, or is it...

This is a big deal for teams already living in @Notion . The "set it once, runs while you sleep" promise is exactly what async-heavy teams need — especially when your docs, tasks, and comms are already in one place. Curious how the agents handle ambiguity, like when a task trigger fires but the context is incomplete. Does it ask for clarification or just make a judgment call? That's usually...