I’m excited to share the release of actix_tower v0.1.3!
Actix Web is famously fast and mature, but many developers find themselves reinventing the wheel for middleware that already exists in the broader tower ecosystem. actix_tower solves this by safely wrapping Send + Sync Tower middleware for use inside Actix Web's !Send thread-local workers.
We just completed a massive overhaul focused on enterprise-grade reliability and performance:
Performance: Bridging overhead is mathematically constrained. We replaced all Box::pin instances on the hot path with stack-allocated, static-dispatch futures.
Reliability: We built a 70+ automated test suite spanning Security & Hardening, Concurrency & Edge Cases, Performance & Regressions, and Stress & Chaos engineering.
Stability: Fully guarded by cargo-semver-checks and an extensive CI matrix testing up to Edition 2024 on Nightly.
If you are building microservices in Actix and want to leverage tower-http (Tracing, Timeouts, etc.) seamlessly, I'd love for you to try it out!
Glad to see you are making progress on your project. You already created an announcement topic for actix_tower though. It would be nice if you could post any follow-ups, like news about new releases, to that topic instead of creating a new one for every update, which could be seen as a bit spammy.
Thanks for the heads up! You're totally right—I'll keep all future updates and release news contained within this thread to keep the forum tidy. I really appreciate the feedback!
It seems like there has been a big uptick in AI code on this forum. I don't have a problem with it in principle but it I would like to know if I'm interacting with a human and how the code was generated. If it's AI, share your prompts and what model you used!
About 90% of the code for actix_tower was written and engineered by me.
I do use AI as an engineering assistant—for brainstorming, reviewing architecture, generating test ideas, and challenging design decisions—but not as a replacement for implementation.
For actix_tower, I designed the architecture, implemented the bridge between Actix Web and the Tower ecosystem, wrote the core implementation, debugged the concurrency issues, and iterated on the design myself. AI was primarily useful as a reviewer and adversarial tester to help identify edge cases and validate correctness after completion of project.
The same workflow applies to my other long-term projects, including ******* (privacy-first analytics platform), ******* (MRI of backends), and my trading infrastructure. AI helps me think, but the engineering decisions, implementation, debugging, and maintenance are my responsibility.
I think AI is a powerful tool, but shipping and maintaining production-quality infrastructure still requires understanding the code well enough to design it, review it critically, fix bugs, and support it over time.
Google helps me alot. I am using Gemini 3.1 Pro (High).
,