Hey everyone 👋
Over the past few weeks, I've been building SmoothAPI, an open-source library for TypeScript and Python that helps applications handle unreliable APIs more gracefully.
The idea came from repeatedly writing the same retry and failure-handling logic across different projects. Every time an API became slow, rate-limited, or temporarily unavailable, I'd end up implementing retries, backoff logic, and fallback behavior from scratch.
So I decided to build a reusable solution around those patterns.
Current features include:
-Exponential backoff with full jitter
-Circuit breakers
-Intelligent retries
-Graceful fallbacksTypeScript and Python support
I'm now working on improving the ecosystem around it with examples, documentation, and community contributions.
Would love feedback from other developers:
- What's the biggest pain point you've faced when integrating external APIs?
- Are there any resilience features you'd expect from a library like this?
GitHub: https://github.com/aryansharma48/smoothapi
Website: https://smoothapi.org
Interesting build.
What I've seen trip teams up isn't usually the API failure itself.
It's discovering which failures should be hidden from users and which failures should be visible immediately.
That's a great point.
One of the things I'm trying to balance with SmoothAPI is giving developers sensible defaults while still letting them decide when a failure should surface immediately versus when it should be retried or handled through a fallback.
The goal is to make common failure scenarios easier to manage without taking control away from the application itself.
Appreciate the insight.
That's the part I'd watch closely.
A lot of products start by helping teams handle failures better and end up discovering they're actually making a much bigger choice about what responsibility belongs to the platform versus the developer.
Interesting area.