Nango

Not everything needs to run right now…but it does need to run. Until recently, Nango Actions were synchronous and immediate. Great for fast workflows, but not ideal for: - Pushing large volumes of data - Updating external systems where consistency matters over speed - Ensuring retry logic in case of failures 💡 That’s where Async Actions come in! With this new capability, you can now: - Trigger a long-running task - Let Nango handle queuing, retries, and delayed execution - Poll for results when ready, no more holding open connections Think of it as reliable delivery for integration actions, without the ops overhead. Perfect for: - Monthly invoice pushes - CRM state syncing - Any “this must happen, even if it’s not immediate” use case Learn more about async actions in our docs (link in 1st comment)

"Just connect the AI to the API!" Famous last words. Today's APIs aren’t built for AI agents. And agents aren’t built to robustly handle: ➡️ OAuth flows ➡️ Schema quirks ➡️ Rate limits and much more.. That’s why AI agents need more than a prompt. They need infrastructure (yes, going beyond MCPs). We’re writing more on this soon. Stay tuned!

Integrations are no longer ‘nice to have’. They’re now your competitive edge. Why? The old way of legacy SaaS: Integrations help with adoption and retention. The new way of AI-native SaaS: Integrations DEFINE what your product can do. A sales AI agent is only as good as its access to CRM, email, and call data. A recruiting copilot is only useful if it can execute actions across relevant platforms. In an AI-first world, integrations aren’t secondary features ➡️ they’re the surface area of intelligence.

New team member alert 🔔 Welcome Rune to the Nango team as Head of Growth! He’s seen what it takes to scale... from joining Datadog early to leading GTM at several infra-focused devtool startups Rune’s deep experience with developer audiences and ABM strategy is a perfect match for our next chapter ✨ Can’t wait to see where we go from here!

AI-native IDEs are only as good as the context you feed them. That’s why we’ve built Nango MCP server for Cursor. Here's what this means for you: 🔎 Search and understand Nango docs directly inside Cursor 💡 Get real-time integration guidance in your editor ✅ Follow best practices for implementing Nango 🔐 Set up OAuth + API integrations with AI assistance It’s like having a Nango engineer in your IDE 24/7. 🤝 Check out Nango MCP docs (link in comment)

Our small but mighty engineering team is on FIRE🔥 Here's what the team shipped in April: ✅ 22 new APIs (ADP, Azure DevOps, Workday OAuth & more) ✅ 16 new pre-built integrations (Gong, BambooHR, Microsoft Teams…) ✅ Environment management, now directly in the dashboard ✅ Usage-based pricing starting at $50/month that scales as you scale ✅ Way faster log search ✅ Enterprise self-hosting (ECS + Terraform) More exciting updates coming this month, stay tuned!

Wait…failure is fine? One thing we’ve learned helping companies build AI integrations at Nango: Failure is fine, as long as you have robust systems. We’ve seen AI agents hit ~30% error rate on API calls…and still deliver great results. That’s because AI can reason with error logs and retry effectively. But they need help. They forget context. They misread outputs. The fix? Design your integrations for: - Structured inputs and outputs - Type safety - Well-scoped error handling and logging A structured integrations framework goes a LONG way when building AI integrations. It will help you reduce AI credits consumed, while delivering robust integrations at scale.

The biggest integration regret we hear? ‘We didn’t realize how limited we’d be.’ Before entrusting your customer integrations to an embedded iPaaS, consider this: 1️⃣ No-code isn’t enough: Most platforms promise that anyone can set up integrations. But real use cases still demand deep technical understanding and developer involvement. 2️⃣ Prebuilt doesn’t mean plug-and-play: Prebuilt connectors offered by embedded iPaaS rarely fit your product’s specific needs out of the box. Customization needs often mean extra work, not less. 3️⃣ Not built for developers: The abstractions and tools offered by embedded iPaaS don’t match what developers need for reliable, scalable, and maintainable integrations. Critical features like type safety, continuous sync, or two-way data flow are usually missing. 4️⃣ Scaling gets messy: What works for a handful of users often breaks when your product grows. Most embedded iPaaS solutions hit performance limits long before you expect. If integrations are core to your product, it’s worth thinking carefully about the trade-offs up front. Read our blog post for a full break down (link in 1st comment 🔗)