An offline coding agent for local models. One command sets up the best model your GPU can run, and on 24 GB+ RTX 30/40/50-series cards Qwen3.8-27B runs up to 1.9x faster than llama.cpp.
- Recommended by
- 1
- GitHub stars
- 130
- Licence
- AGPL-3.0
- Language
- Python
- Last commit
- 1 Nov 2026
What people said
“>Qwen 3.8 27B Q4 with 64k, 3090: ~40tok/sec Try more like 70-130 tok/sec.. And also Qwen models famously chase benchmarks. They are good for writing code, but fall behind in other areas. Gemma4:31B is 6gb larger and do…”
Not software, abandoned or listed by mistake? Report this listing