Leo

Congrats on the launch! This resonates with me, Raymond. I've been deep in the knowledge management space and the thing that kills most of these tools is they all eventually try to become Notion. Then nobody knows what the product actually does anymore. The voice-to-docs angle is what caught my eye though. Getting a clean transcript is the easy part now. The hard part? Figuring out what from a...

Congrats on your 4th launch! One thing I've been wrestling with on the MCP side: The protocol is stateless but browser sessions obviously aren't. When someone's agent logs into a portal, clicks through three pages, and pulls invoices, how are you keeping that session alive between tool calls? Are you pinning to a browser instance on your end? That stateless-protocol-meets-stateful-browser...

Here's something most people don't realize ... when you copy formatted text, macOS doesn't store one thing. It stuffs multiple versions into the pasteboard at once: plain text, HTML, RTF, sometimes more. Then every app you paste into just silently picks whichever version it wants. Google Docs grabs the HTML. Slack ignores it and takes plain text. Email clients want inline-styled HTML. That's...

I've been building with Claude Code and the difference between a well-written skill/instruction set and a mediocre one is night and day. The ElevenLabs case study is a compelling proof point. Most people are still treating agent instructions as an afterthought, just a markdown file in the repo. The idea that you can actually evaluate and iterate on them like any other piece of software makes a...

Congrats on the launch! MCP discovery is genuinely one of the biggest friction points right now. There are hundreds of servers popping up but being able to try a server in the browser before wiring it into your config is a great call. I built an MCP server for my product and the hardest part isn't the integration ... it's getting people to discover it exists and understand what it actually does...

Charlie, this hits close to home. I've been building developer tools that sit between AI conversations and the real world, and the "knowledge dies in the chat window" problem is one of the most underappreciated bottlenecks in AI-assisted workflows right now. Your Cloud Run example is perfect. I've had the exact same experience. An agent spends 45 minutes figuring out a Supabase migration edge...

The gap between "this open source tool is incredible" and "I actually run it in production" is where most developer tools lose people. SSH, environment config, dependency juggling ... that's not the work anyone signed up for. I've been through this exact cycle building my own infrastructure. You spend a weekend getting something deployed, it works great, then three weeks later a dependency...

This is one of those tools that sounds like a fun gimmick until you actually use it and realize it solves a real workflow problem. I run Claude Code sessions constantly, and the silent terminal problem is real. I'll kick off a task, switch to another window to do something else, and then 15 minutes later realize the agent has been waiting for approval the whole time. That context switch tax...

The Cursor analogy is apt but I'm curious how it translates to writing in practice. With Cursor, the interaction model is clear: tab-complete, inline diff, accept/reject. For code that works because changes are structural and self-contained. Professional writing is messier. If I'm halfway through a 10-page proposal and I ask the agent to "make the competitive analysis more specific," how does...

Great product. This solves a problem I face daily. Like many, when sessions compact I lose so much decision making context. Question: Can this auto-save? Claude supports hooks, so maybe a pre-compaction trigger that auto publishes to Claude Bin? Personally, my pain point isn't necessarily forgetting to share a session it's just losing it before I get the chance.

Nice benchmark numbers. My concern is always the gap between benchmarks and the actual developer experience. I use Claude primarily for coding because, from my personal experience, it follows instructions pretty closely (though there's always room for improvement). For me, Gemini has historically been frustrating for me, inserting comments and refactoring code I didn't ask it to do. Would love...

Congrats on the launch! This is the kind of tool I wish I'd found years ago. I've been using karabiner-elements to remap keys but the config files are a pain. The double tap to restore actual caps Lock is a nice touch. Question ... does the H/J/K/L navigation work inside any app including terminals and code editors? I'd love to use it in VS Code but I'm wondering if it conflicts with...

Congrats on the launch! As someone building a product with AI integration, I'm curious about your conversational UX decisions. When a user says "find me 50 SaaS CTOs in NYC who recently raised Series B," how much back-and-forth refinement typically happens before the list is useful? Do most users get what they need in one prompt, or is it more of an iterative conversation?