Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Zig INferenCe Engine — Local LLM inference on AMD GPUs and Apple Silicon
| Date | Stars |
|---|---|
| 2026-07-31 | 491 |
| 2026-08-07 | 494 |
| 2026-08-15 | 498 |
| 2026-08-18 | 499 |
| 2026-08-19 | 501 |
| 2026-08-20 | 503 |
| 2026-08-21 | 502 |
| 2026-08-22 | 503 |
| 2026-08-23 | 504 |
| 2026-08-24 | 505 |
| 2026-08-26 | 506 |
| 2026-08-27 | 505 |
| 2026-08-31 | 507 |
| 2026-09-01 | 506 |
| 2026-09-02 | 509 |
| 2026-09-03 | 510 |
| 2026-09-04 | 511 |
| 2026-09-05 | 512 |
| 2026-09-07 | 513 |
| 2026-09-09 | 514 |
| 2026-09-19 | 514 |
| 2026-09-20 | 513 |
| 2026-09-23 | 514 |
| 2026-09-25 | 515 |
| 2026-09-26 | 515 |
| 2026-09-29 | 516 |
| 2026-10-01 | 518 |
| 2026-10-06 | 520 |
| 2026-10-13 | 521 |
| 2026-10-15 | 522 |
| 2026-10-20 | 523 |
| 2026-10-25 | 525 |
| 2026-10-27 | 525 |
| 2026-10-31 | 526 |
| 2026-11-01 | 527 |
| 2026-11-02 | 528 |
| 2026-11-04 | 528 |
Today
— stars today
This week
+3 stars this week
This month
+10 stars this month
Momentum
3.0
growth rate 0.57%/day
<p align="center"> <img src="assets/zinc_trademark_new.png" alt="ZINC" width="240"> </p> # ZINC Run GGUF language models on your own GPU. ZINC gives you a command line, browser chat, and an OpenAI-compatible API in one native binary. [](https://github.com/zolotukhin/zinc/actions/workflows/test.yml) [](.zig-version) [](LICENSE) [Get started](https://zolotukhin.ai/zinc/docs/getting-started/) · [Documentation](https://zolotukhin.ai/zinc/docs/) · [Benchmarks](https://zolotukhin.ai/zinc/benchmarks/) · [Discord](https://discord.gg/QRUgWH2aGV) ## Get running For published archives, runtime requirements and verified installation/rollback, see [binary releases](https://zolotukhin.ai/zinc/docs/binary-releases/) and [GitHub Releases](https://github.com/zolotukhin/zinc/releases). The 0.1.0 binary pipeline is undergoing [qualification](docs/releases/0.1.0-plan.md). The commands below build from source. Install the [prerequisites for your GPU](https://zolotukhin.ai/zinc/docs/getting-started/#install-dependencies) first. The default build uses **Vulkan on Linux** and **Metal on macOS**. The build wrapper downloads and verifies **Zig 0.17.0** when needed; you do not need to replace your system compiler. Python 3 is needed for that initial compiler download. ```bash git clone https://github.com/zolotukhin/zinc.git cd zinc ./scripts/zig.sh build -Doptimize=ReleaseFast ``` On AMD with Vulkan, set `export RADV_PERFTEST=coop_matrix` in the same terminal before the GPU check or inference. Then check your GPU, download a model, and open the chat: ```bash ./zig-out/bin/zinc --check ./zig-out/bin/zinc model pull qwen35-9b-q4k-m ./zig-out/bin/zinc chat --model-id qwen35-9b-q4k-m ``` The example uses Qwen 3.5 9B. Check the [memory requirements](https://zolotukhin.ai/zinc/docs/hardware-requirements/) before downloading, or run `./zig-out/bin/zinc model list` to browse the catalog. You can also load an existing file with `-m /path/to/model.gguf`. ## Choose your GPU path | Hardware | Backend | Setup | | -------------------------- | ------------------- | ---------------------------------------------------------------------------------------------------------- | | AMD Radeon RDNA3 / RDNA4 | Vulkan or ROCm/HIP | [Vulkan](https://zolotukhin.ai/zinc/docs/getting-started/) · [ROCm](https://zolotukhin.ai/zinc/docs/rocm/) | | Apple Silicon | Metal | [macOS](https://zolotukhin.ai/zinc/docs/getting-started/#macos-apple-silicon) | | Intel Arc Xe2 / Battlemage | Vulkan | [Intel guide](https://zolotukhin.ai/zinc/docs/intel-gpu-reference/) | | NVIDIA RTX | CUDA · experimental | [Requirements](https://zolotukhin.ai/zinc/docs/hardware-requirements/) | For AMD with a working ROCm installation, choose the HIP backend at build time: ```bash ROCM_PATH=/opt/rocm ./scripts/zig.sh build -Dbackend=rocm -Doptimize=ReleaseFast ROCR_VISIBLE_DEVICES=0 ./zig-out/bin/zinc --check ``` Model coverage and performance vary by backend. The [hardware guide](https://zolotukhin.ai/zinc/docs/hardware-requirements/) lists validated cards, operating systems, drivers, and current limitations. ## Use it your way **Browser chat** in the default Vulkan or Metal build starts the UI and local API together: ```bash ./zig-out/bin/zinc chat --model-id qwen35-9b-q4k-m ``` **One prompt** from the terminal: ```bash ./zig-out/bin/zinc --model-id qwen35-9b-q4k-m --chat --prompt "Explain a ring buffer." ``` **An existing app** can connect to `http://localhost:9090/v1` while the serve
Excerpt of 7,720 characters
Read on GitHub
4.7k
24
22
8
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:2a0022cbc99a3e81, topic:transformer, topic:gpt
matched fp:2a0022cbc99a3e81, topic:gpu
matched fp:2a0022cbc99a3e81, topic:pytorch