# AI use cases for llmfit (AlexsJones/llmfit)

Repo: [AlexsJones/llmfit](https://github.com/AlexsJones/llmfit) | Use cases: 3 | Page: https://dedupedai.com/use-cases/alexsjones-llmfit/

llmfit inspects your CPU, RAM, GPU and VRAM and recommends which open-source LLMs and quantizations your hardware can comfortably run, with a TUI, JSON output and real benchmark runs.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Find which open models your machine can run: https://dedupedai.com/use-cases/alexsjones-llmfit/find-models-your-hardware-can-run/
- Get model recommendations as JSON for scripts: https://dedupedai.com/use-cases/alexsjones-llmfit/pick-coding-model-from-json/
- Measure real tokens per second on your hardware: https://dedupedai.com/use-cases/alexsjones-llmfit/benchmark-model-speed-on-your-machine/

## Which LLM can my PC run? Check your hardware with llmfit

Repo: [AlexsJones/llmfit](https://github.com/AlexsJones/llmfit) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README

Run llmfit to see your hardware and a ranked list of models that fit.

**Who it is for:** A developer about to download a local model who does not know what fits.

**The problem:** Downloading a model that will not fit wastes hours.

**The outcome:** A ranked list of models and quantizations for your hardware.

### Steps

1. Install with `brew install llmfit` or `uv tool install -U llmfit`
2. Run `llmfit` for the interactive TUI
3. Browse models ranked for your hardware
4. Pick one and download it

### Prompt to paste into your agent

```text
Run AlexsJones/llmfit and list the top five recommended models for my hardware with their quantization. Confirm the detected GPU and RAM match my system.
```

### Caveat

Recommendations are estimates until replaced by real benchmark runs.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/alexsjones-llmfit/find-models-your-hardware-can-run/

## How to get model recommendations as JSON for scripts with llmfit

Repo: [AlexsJones/llmfit](https://github.com/AlexsJones/llmfit) | Difficulty: Intermediate | Time: 15 minutes | Basis: From the README

Use recommend with --json and a use case to feed other tools.

**Who it is for:** A developer scripting local model setup across machines.

**The problem:** Manual model choice does not scale across machines.

**The outcome:** JSON recommendations filtered by use case.

### Steps

1. Install llmfit
2. Run `llmfit recommend --json`
3. Filter by use case, for example coding, in Docker or podman
4. Parse the model names with jq

### Prompt to paste into your agent

```text
Run `llmfit recommend --json` and return the top three model names for coding on this machine.
```

### Caveat

JSON fields can change between versions; pin a version for scripts.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/alexsjones-llmfit/pick-coding-model-from-json/

## How to measure real tokens per second on your hardware with llmfit

Repo: [AlexsJones/llmfit](https://github.com/AlexsJones/llmfit) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Download, serve and benchmark a model from the TUI, then optionally share results.

**Who it is for:** A hobbyist comparing models on a home GPU.

**The problem:** Published speed numbers rarely match your hardware.

**The outcome:** Measured tok/s saved locally that replace estimates.

### Steps

1. Open the llmfit TUI
2. Choose a model to download and serve
3. Run the benchmark
4. Optionally contribute results as a PR from the TUI

### Prompt to paste into your agent

```text
Use llmfit to benchmark <model> on this machine and report measured tokens per second versus the estimate.
```

### Caveat

Benchmarks download full models, which can be many gigabytes.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/alexsjones-llmfit/benchmark-model-speed-on-your-machine/