# AI Inference engine built in Zig for Mac's, MLX-Serve

**URL:** <https://ziggit.dev/t/ai-inference-engine-built-in-zig-for-macs-mlx-serve/15498>\
**Category:** Showcase\
**Tags:** llm\
**Created:** [May 13, 2026, 1:48pm UTC](https://ziggit.dev/t/ai-inference-engine-built-in-zig-for-macs-mlx-serve/15498 "2026-05-13T13:48:27Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![ddalcu](https://ziggit.dev/user_avatar/ziggit.dev/ddalcu/32/8676_2.png) [@ddalcu](https://ziggit.dev/u/ddalcu)\
**Post date:** [May 13, 2026, 1:48pm UTC](https://ziggit.dev/t/ai-inference-engine-built-in-zig-for-macs-mlx-serve/15498/1 "2026-05-13T13:48:27Z")

</div>

Hey all, built a native zig mlx ai inference engine for MacOS & Apple Silicon.

Zig is slightly faster than Python obviously, however most AI inference is memory is the bottleneck, so also added some speculative decoding to increase speed, depending on workload types, more charts, documentation and full sourcecode available at my github repo. [[https://github.com/ddalcu/mlx-serve/](https://github.com/ddalcu/mlx-serve/)

Dev branch: working on adding DeepSeek V4 Flash support, which is proving to be difficult, but made really good progress. Had to quantize it down to 2bit so it can fit on my 128gb Mac.

Main Branch: Supports most MLX models, Qwen 3.6, Gemma, MoE, MTP/Assist models, Dense models, etc.  
UI: Swift native code that launches the Zig server, to manage model downloads, chat UI, TaskTray, Launch coding agents if you have those installed, and has a optional built in Agent also + MCP.

 ![Screenshot 2026-05-13 at 9.46.59 AM](https://ziggit.dev/uploads/default/original/2X/f/fcfb9da2fa1245e5147f255128dafc3bb716b455.jpeg)

---

<div class="post-metadata">

**Author:** ![ddalcu](https://ziggit.dev/user_avatar/ziggit.dev/ddalcu/32/8676_2.png) [@ddalcu](https://ziggit.dev/u/ddalcu)\
**Post date:** [July 11, 2026, 7:56pm UTC](https://ziggit.dev/t/ai-inference-engine-built-in-zig-for-macs-mlx-serve/15498/3 "2026-07-11T19:56:51Z")

</div>

Added a few new features to the AI Inference engine, it can now do:  
Music Generation  
Voice Cloning  
Video Generation  
Image Generation

All built in Zig ! All 100% Local ! Please star the GitHub repo if you like, PR’s are welcome !

 ![Screenshot 2026-07-11 at 3.53.20 PM](https://ziggit.dev/uploads/default/original/2X/d/ddfc2c2491261bd913547d3d8e63b03a931b0a23.jpeg)