Big news: We've supercharged @Maple AI reliability & performance with PlanetScale! 🚀
Zero downtime migration, 45% faster API responses, and game-changing observability for our encrypted workloads.
At OpenSecret, we power truly private AI chat (Maple AI) using AWS Nitro Enclaves. User data is always encrypted - even we can't see it. This makes database reliability and deep observability non-negotiable. Guesswork isn't an option.
Our previous provider struggled with our always-on demands. During a critical week, we suffered hours of downtime across 4 incidents. Cold starts hurt UX, usage-based pricing punished growth, and we had zero performance visibility.
PlanetScale changed everything:
✅ Predictable pricing: $156 | 130,261 sats | 130,357 sats/mo for 4 DBs + 8 replicas (vs $250 | 208,751 sats | 208,906 sats for less elsewhere!)
✅ Instant p99 latency (1.0ms!) & query insights
✅ Built-in failover & 100% uptime since migration
✅ Migrated with ZERO downtime using their tools
The performance leap shocked us:
• Slowest API endpoint latency dropped 45% (550ms → 300ms)
• Finally see which queries need optimization
• Insights work perfectly on encrypted data - critical since we debug "blind"
Cost savings? 38% less while getting 3X the infrastructure. But the real win is confidence: automatic failover, replicas always running, and metrics that actually help us optimize.
We build confidential computing; PlanetScale masters databases. This focus lets us innovate faster on our confidential app platform for apps like Maple AI while knowing our foundation is solid. Branching, sharding, and Metal tier await as we scale! 🙌
You can read the full technical writeup on our blog: 
At OpenSecret, we power truly private AI chat (Maple AI) using AWS Nitro Enclaves. User data is always encrypted - even we can't see it. This makes database reliability and deep observability non-negotiable. Guesswork isn't an option.
Our previous provider struggled with our always-on demands. During a critical week, we suffered hours of downtime across 4 incidents. Cold starts hurt UX, usage-based pricing punished growth, and we had zero performance visibility.
PlanetScale changed everything:
✅ Predictable pricing: $156 | 130,261 sats | 130,357 sats/mo for 4 DBs + 8 replicas (vs $250 | 208,751 sats | 208,906 sats for less elsewhere!)
✅ Instant p99 latency (1.0ms!) & query insights
✅ Built-in failover & 100% uptime since migration
✅ Migrated with ZERO downtime using their tools
The performance leap shocked us:
• Slowest API endpoint latency dropped 45% (550ms → 300ms)
• Finally see which queries need optimization
• Insights work perfectly on encrypted data - critical since we debug "blind"
Cost savings? 38% less while getting 3X the infrastructure. But the real win is confidence: automatic failover, replicas always running, and metrics that actually help us optimize.
We build confidential computing; PlanetScale masters databases. This focus lets us innovate faster on our confidential app platform for apps like Maple AI while knowing our foundation is solid. Branching, sharding, and Metal tier await as we scale! 🙌
You can read the full technical writeup on our blog: OpenSecret
Why We Migrated from Neon to PlanetScale
In May 2025, during the same week Neon announced their acquisition, our databases went down four times. For hours. Database spin-ups, their entire ...
Developers today live in an era where user data either sits in plain text in databases or is encrypted in a way that any admin or server with access keys can get ahold of and process their data in any way. This has led to many data leaks in the past.
Current attempts at doing E2EE have burdened the responsibility onto the user in ways where any end-user mistake can be drastic for them. Alternative approaches like iCloud or Google Drive encryption mean that data is often not available to access on all user devices.
Secure enclaves provide a great middle ground. They allow you to verify the code that is running while protecting sensitive data that runs inside a server's protected memory with private keys that only the verified enclave can access.
We built our developer platform on top of Secure Enclaves, elevating the current dev experience into a security-by-design approach on day 1. Our offering provides the developer experience they expect while offering their users great security out of the box.
We offer a powerful combination of user authentication, private key management, encrypted data sync, and private AI inference. This entire stack runs on secure enclaves, protecting users no matter how they use your app.
Our open source dev platform is fully verifiable through reproducible builds and attestation verifications against those builds, so you and your users know immediately that the OpenSecret backend is running exactly what they expect while being protected in memory.
We built Maple AI on top of our platform as a proof of concept, but you can build any app that you can build today. Consumer productivity apps, health tracking apps, note tracking, fintech, and many more can be built on this to protect user data around every corner.
Our developer preview environment is available today. We would love to have you get started building on this secure platform. Join our Discord or DM to get access.
We believe the future of servers will be on top of confidential computing. We're making this accessible to any dev today.
Read more detail in our blog post on this topic. Keep reading this thread for the highlights.
