llml LLM Launcher

New profiles added regularly — watch the llml repo ↗ to get notified.

LLM Launcher

Stop reconstructing llama-server flags from shell history.

llml finds your local models, detects your runtimes, and launches them with a saved profile. Pick a model, pick a profile, press R.

Works alongside llama.cpp, Ollama, vLLM, and KoboldCpp. Your backend stays exactly as it is.

$ curl -fsSL https://llml.dev/install.sh | sh
Alternatives: Homebrew cask (macOS), Scoop/Winget (Windows), or Go. See Install instructions ↗
Install llml ↗ Browse profiles →
llml terminal UI showing model scan, runtime selection, and profile launch workflow

What llml does

the loop
01

Scan

llml finds every local model — GGUF files, safetensors, Hugging Face cache — and lists the runtimes on your machine.

02

Pick

Choose a model, a runtime, and a saved profile. The generated launch command is shown before execution — no surprises.

03

Launch

Press R. One keypress, and the right command runs against the right model on the right backend.

Read more about profiles →

Don't start from scratch.

Import the exact config someone already tuned for your model and hardware — one command, it runs. Every profile is a TOML file with args, env vars, and hardware metadata. The catalog matches them to your machine.

Profiles are starting points, not guarantees. Hardware differs — expect to adapt. But starting from someone's working config beats starting from an empty terminal.

Find a profile for your machine →
$ llml import https://llml.dev/profiles/Qwen3.6-enable-thinking.toml --activate

llml export writes a portable TOML. The catalog is where those get shared. llml import pulls one back. Export → share → import → run.

Recently updated

Browse all 16 profiles →
llama.cpp community

gemma-4-26B-A4B-31B

26B-A4B: 16-18 GB (4-bit), 28-30 GB (8-bit), 52 GB (BF16/FP16).

Mixed Cross-platform General
llama.cpp community

gemma-4-26B-A4B-31B-image

26B-A4B: 16-18 GB (4-bit), 28-30 GB (8-bit), 52 GB (BF16/FP16).

Mixed Cross-platform General
llama.cpp community

gemma-4-26B-A4B-31B-thinking

26B-A4B: 16-18 GB (4-bit), 28-30 GB (8-bit), 52 GB (BF16/FP16).

Mixed Cross-platform General
llama.cpp community

gemma-4-26B-A4B-31B-thinking-image

26B-A4B: 16-18 GB (4-bit), 28-30 GB (8-bit), 52 GB (BF16/FP16).

Mixed Cross-platform General
llama.cpp community

gemma-4-E2B-E4B-12B

E2B: 4 GB (4-bit) or 5-8 GB (8-bit), phone/edge.

Mixed Cross-platform General
llama.cpp community

gemma-4-E2B-E4B-12B-image-audio

E2B: 4 GB (4-bit) or 5-8 GB (8-bit), phone/edge.

Mixed Cross-platform General

What makes this different

Importable recipes, not fit data

Not a list of what fits. Not per-backend defaults. The exact config someone tuned for your model and hardware — imported and run with one command.

Portable, not platform-locked

Profiles are TOML files with a documented schema. They live in GitHub, not a service. The catalog is a thin index over real source files.

Reproducible imports

An import is one shell command. The same TOML produces the same args and env on every machine — no hidden web of preferences.

Got a profile that just works?

Share the args your machine and model converged on. PR-only. No accounts.

Read the format Open a PR ↗