Ruby - Ask better questions, live on every call
by•
Most people walk out of an important call thinking "I should have asked that." Ruby makes sure you don't.
She listens in real time and slides you the question worth asking — through a small pill only you can see, live as the conversation unfolds.
→ Before: she helps you think through your goal and what to listen for.
→ During: she nudges you with the sharper question, at exactly the right moment.
→ After: she writes the recap, so you don't have to.
Replies
Grass
The invisible pill UI is the right call, the moment it becomes visible to others it turns into a crutch and kills authenticity. What I'm curious about: does Ruby learn from the calls that went well? Like, if I closed a deal after asking a specific sequence of questions, does it start surfacing similar patterns in future calls?
Bookshelf – Turn Newsletters into GPTs
@romeo_ciobanu Not yet, but that's a great idea! Today you can manually add instructions to Ruby's memory to personalize her for you
@sahil_kathpal1 That's actually a smart workaround for now, manually curating the memory means you're being intentional about what worked, rather than hoping the AI figures it out on its own. Would love to see the automatic pattern learning when it ships. Congrats on the launch, rooting for you!
Bookshelf – Turn Newsletters into GPTs
@romeo_ciobanu Exactly! And thanks! Would love to hear what you think once you've tried it out :)
Grass
@luki_notlowkey That's actually the problem we're trying to solve 😄
We're still early, so I don't have a legendary example yet. But even in our own calls, Ruby has been good at catching the follow-up questions we'd otherwise think of after hanging up.
The value isn't that it asks magical questions. It's that it helps you ask the obvious next question at the moment it matters.
Bookshelf – Turn Newsletters into GPTs
@luki_notlowkey It did a great one for me for another of my products.
In an early pilot discussion, it caught the customer mentioning they were nervous to deploy our tool with their live customers. Ruby nudged me to go deeper on that and prodded for specifics. Led to them ackowledging size was the concern and they'd be okay piloting our tool with a smaller sized customer. We closed the pilot!
Bookshelf – Turn Newsletters into GPTs
One thing worth flagging for anyone getting an error on first launch 👇
You'll need Claude Code installed in your terminal for Ruby to work, not just the Claude desktop app. They're two separate things, and the app alone won't connect.
If you've hit an error during your first prep chat, that's almost always why. Install Claude Code in your terminal and you'll be set.
Happy to help if anyone's still stuck 🙌
Does Ruby run its real-time transcription and suggestions locally, or is call audio sent to the cloud? Curious how that works for calls under NDA or with privacy-sensitive interviewees.
Bookshelf – Turn Newsletters into GPTs
@nikita_zaro the transcription does use a third-party cloud api. No audio is ever stored on our end though, simply passes through a relay.
Ruby looks like the kind of tool that’s useful because it stays simple. I like products that solve one clear problem without making you learn a whole new workflow first.
Bookshelf – Turn Newsletters into GPTs
@rick_borduur appreciate that Rick! Would love to hear what you think once you've tried it out :)
Would love a tutorial on how to get started!
Bookshelf – Turn Newsletters into GPTs
@william_ross4 you need a mac and claude code installed. That's it. It'll walk you through everything once you start the app.