@andrewparker Check out @getskydive . We do secrets injection in our egress proxy, but also generally recommend creating multiple agents, each having a different permissions boundary.
@atharvmantr_i@ftcarpe Depends on what sort of telemetry/observability you have in place. We have customer tagging on data like Cloudwatch logs and Sentry, which help our agents determine that a specific customer might be experiencing above error counts.
@ASpittel Honestly, this was a lot of our motivation for @getskydive . Build, iterate, and run your agent in one place. Have it interact with any other services you use.
@atharvmantr_i@ftcarpe Definitely. We use agents internally to proactively alert us about customers having issues, and then do reach out to them. Also follow-up once things are resolved.
@mattpocockuk Made this jump months ago and do it all through @getskydive . Does help that anytime I find something that irks me in the platform, I just fix it. ๐
@1arnavsurve Usually, my agent just chats right back. Low key, I think Linus and Nephew are pals. I'm guessing they have convos we don't even know about.
what if AI teammates (Grok Bot, Hermes, etc.) lived everywhere?
Introducing Skydive by Anything.
- omnichannel (Slack, email, iMessage, CLI)
- cloud native (always on, works while you sleep)
- model agnostic (Fable, GPT, Deepseek, more)
everyone gets full access from day one
let's fly. ๐ช
Mocha is winding down.
After two years, weโve made the decision to sunset the product. We started with a bet that anyone with an idea should be able to build actual, useful, real working apps with auth, databases, backends, and one-click deploys. We believed the people locked out of building software had some of the best ideas of anyone.
And for a lot of you, it worked. Real small business owners and creators built actually useful software with Mocha: customer portals, booking systems, internal tools, side projects that turned into businesses. People who'd never written a line of code, running their work on software they built themselves. That part was real.
But the business underneath had its challenges. A combination of high user acquisition costs driven by competition, expensive unit economics due to AI token requirements of the product, and high support costs all compounded into making the business capital intensive. We weren't able to raise the capital to withstand this over the long run. Closing is the right call for our team and our investors.
The best part about our journey was you, our users. We care deeply about our users, so we spent real time finding the right home for your apps โ and we found one in @anything. They share our philosophy: integrated, opinionated, built for people who describe outcomes and want working software. As a bonus, they support mobile apps, something many of you have been asking us for. Mocha apps will be in good hands there.
They've built a one-click migration that ports your apps over and keeps them live. We've tested it. It works.
Theyโre also going to provide dedicated help to anyone looking to migrate over.
Timeline:
Today: New signups closed. Existing customers, we're here for you.
Through July: Migration support and reminders.
August 1: Mocha's app builder shuts down. Migration and export tools stay open for 30 days after.
Move your apps โ https://t.co/vHsK9S6o0q ($20 in credits to help with post-migration edits) Prefer to export and self-host? Settings โ Export.
To everyone who built something with Mocha: thank you. Watching what you made was the best part of this journey. The era of idea people and personal software isn't over. It's just beginning.
โ Nicholas & Ben