Codex CLI
Lets you hand coding tasks to OpenAI's agent, which edits and runs code on your computer
by OpenAI
llmfit
by Alex JonesCLI toolWrite code
Shows which AI models your computer can run, ranked by fit, speed and quality
The problem it solves
You want to run an AI model on your own computer, but model lists do not tell you which ones your memory and graphics card can handle.
For Local AI hobbyistsDevelopersTechnical makers
- Before
- You guess from model sizes, download several gigabytes, and find out it is too slow or will not load.
- After
- You run one command, see your hardware and a ranked list of models that fit, with speed estimates.
- Stars
- ★ 39k
- License
- MIT
- Cost
- Free
- Platforms
- Mac · Win · Linux
- Level
- Intermediate
Overview
llmfit is a terminal tool that reads your computer's memory, processor and graphics card, then scores hundreds of open AI models for how well they would run on it. It opens an interactive list by default, and also offers plain command output, a web dashboard and a local API. It works with local runtimes such as Ollama and LM Studio. It suits anyone choosing a local model who does not want to download blindly.
One command turns your computer's specs into a ranked list of models that fit, with speed estimates.
Try these first
Start with the first one: it is a small task you can check easily.
- Show the models that fit my computer: llmfit recommendFirst try · read-only
- Save the recommendations as JSON: llmfit recommend --jsonread-only
- Estimate the disk space for keeping three runnable models: llmfit storage --keep 3 --selection largest --jsonread-only
- Open the interactive list and filter by a model name: llmfitread-only
Example prompts written by Agent Flows Hub.
Features
From the creator's documentation.
| Feature | What it does | Runs |
|---|---|---|
| Hardware detection | Reads CPU, memory, graphics card and VRAM, including Apple Silicon | Your computer |
| Model scoring | Scores each model for quality, speed, fit and context length | Your computer |
| Terminal and web views | Interactive terminal list, plain command output and a web dashboard | Your computer |
| Local API | Serves JSON endpoints for your system and the model list | Your network if you expose it |
| Local runtimes | Works with Ollama, llama.cpp, MLX, Docker Model Runner and LM Studio | Your computer |
| Shared benchmarks | Optional: send your measured speeds to the project as a pull request | GitHub |
Good to know
You want to run an AI model on your own computer and need to know which ones will actually work. llmfit detects your hardware and ranks models by how well they fit, so you can pick one before downloading it.
Limits
- On Windows it detects memory and processor; a graphics card is detected only when nvidia-smi is installed.
- Speed figures are estimates until you run a benchmark, which then replaces the estimate with your own measurement.
Notes by Agent Flows Hub (original project name: llmfit), written from the creator's documentation, not from running it ourselves. llmfit README · llmfit platform and GPU support
Stuck setting it up?
Tell us what happened. We take a look first; if it needs paid work, you get an estimate before anything starts.
More to write code
Other cli tools
- MarkItDown Converts PDFs, Word, Excel, PowerPoint and more into Markdown text an AI can read
- OfficeCLI Lets your AI agent create and edit Word, Excel and PowerPoint files without Office
- Open Code Review Reads your Git changes and writes line-by-line review comments with an AI model
- OpenCLI Lets your AI agent use websites through the Chrome browser you are already signed in to
- OpenCommit Writes a clear git commit message for your changes with one short command
- Repomix Packs a whole code project into one file you can hand to an AI chat assistant