Shoehorn is listed as Free. Check https://notactuallytreyanastasio.github.io/shoehorn/ for the current plan limits, as vendors change pricing frequently.

What is Shoehorn best for?

Shoehorn is commonly used for Local inference, Limited VRAM, Model quantization. It supports English.

Shoehorn

Q: What is Shoehorn?

Shoehorn is a code AI tool listed on Use AI Tools. Shoehorn makes a language model fit the memory you actually have. Preset quantizations ignore your hardware: pick one that fits and you either waste quality headroom or discover at load time that it did not fit after all. Shoehorn starts from your available memory, subtracts what inference itself needs, and solves a per-tensor mixed-precision assignment that reportedly uses about 99.99% of the remaining budget. It runs on Mac, Linux, and Windows.

Code⚡ Advanced✅ Global access

Local inferenceLimited VRAMModel quantization

FreeEnglish

Shoehorn makes a language model fit the memory you actually have. Preset quantizations ignore your hardware: pick one that fits and you either waste quality headroom or discover at load time that it did not fit after all. Shoehorn starts from your available memory, subtracts what inference itself needs, and solves a per-tensor mixed-precision assignment that reportedly uses about 99.99% of the remaining budget. It runs on Mac, Linux, and Windows.

Visit Website Compare

Community Rating

Be the first to rate Shoehorn. Your feedback helps other users decide.

🎯 Real-World Use Cases

Fit a Model to Limited VRAM

Quantize to the memory you actually have

Per-Tensor Mixed Precision

Go beyond one-size-fits-all quantization presets

Avoid Load-Time Failures

Stop discovering at load time that a model does not fit

⚖️ Honest Review

✅ Pros

+Starts from your real available memory
+Per-tensor mixed precision instead of a fixed preset
+Reported to use nearly all of the remaining budget

⚠️ Cons

−Open-source project with a small user base
−Quantization still costs some model quality

📚

Guides and comparisons about code tools

Infrastructure tools for AI coding agents

CodeOct 18, 2026