Commands
Pick the command you want the full reference for.
Each page covers what the command is for and when to reach for it, every flag, five real examples, a real captured run, and the errors it actually prints with what to do about them.
Get a model running
- omm searchFind a model across the curated catalog, HuggingFace and ModelScope.
- omm installDownload a model into the hub and link it into every installed runner.
- omm runChat with an installed model — in the terminal for Ollama, or by opening the app for GUI runners.
- omm recommendGet a model suggestion ranked for this machine's hardware, with an offer to install it.
- omm setupRe-run the hardware scan and runner-install checklist, any time.
Manage installed models
- omm importAdopt .gguf files sitting in other apps' model directories into the omm hub.
- omm uninstallRemove a model and clean up its symlinks and manifests. Alias: rm.
- omm listShow every model omm has installed and which runners each is linked into. Alias: ls.
- omm infoShow full detail — repo, version, size, links, run commands — for one installed model.
- omm upgradeRefresh installed models against their source — only re-downloads what's actually changed. Alias: up.
Check hardware & fit
Measure quality & speed
Fix & maintain
- omm linkRe-verify and repair every installed model's runner links, or link into a custom directory.
- omm cleanupClean up leftover partial downloads and broken runner symlinks in one pass — no flags needed.
- omm doctorDiagnose the omm install and Ollama links, read-only — no flags needed.
- omm engine installInstall one local AI runner program directly, skipping the setup checklist.
- omm logRead the local run log: what omm ran, when, and whether it worked.
Configure omm
Every command the CLI has
Taken straight from omm's own help output, so a command added to the CLI shows up here without anyone editing this page.
- omm benchmarkMeasure a small reproducible quality pack and decode speed.
- omm cleanupClean up leftover partial downloads and broken runner symlinks.
- omm contributeBenchmark models in a loop to improve `omm recommend`.
- omm doctorDiagnose the OMM install and Ollama links without changing state.
- omm engine5 sub-commandsInspect, install, update, and remove local AI runner programs.
- omm exportExport a hub model to `destination` for deployment or backup: a hard link when possible, otherwise a real copy. Never a symlink, so the exported file keeps working after `omm uninstall` or on another machine. Not tracked in the registry - uninstalling the source model never touches an exported copy. Also writes a provenance/checksum manifest sidecar next to it, so `omm import` on another machine (including an air-gapped one) can restore the source repo, version, and install date instead of treating the file as an anonymous import.
- omm fitShow whether a model fits this PC's memory right now - installed or not - as a bar over what other apps use, the OS reserve, and the install cap.
- omm helpShow help, same as --help.
- omm importAdopt .gguf files from other local AI apps into the omm hub.
- omm infoShow what a model is - source repo, version, size and linked-program run commands once installed, or author, downloads, license and architecture for a search result. `omm fit` is what tells you whether it runs here.
- omm installDownload a model into the central hub and link it into installed engines.
- omm linkLink models into every supported app, an arbitrary directory, or both scoped to a chosen subset of models.
- omm listalias: lsShow models installed via omm and their linked status.
- omm logShow the local run log (~/.omm/logs/history.log).
- omm pinMark an installed model to be archived before its next `omm install --force`, so `omm rollback` can undo that reinstall afterwards. Nothing is copied yet - GGUFs are large, so the archive is only made right before the reinstall actually replaces the file, not at pin time. `omm install --force` still replaces a pinned model same as any other; pin only decides whether the version it replaces is kept.
- omm recommendSuggest a model to install for this hardware.
- omm rollbackRestore a pinned model's archived version in place of the one currently installed. The version it replaces takes over the archive slot, so running `omm rollback` again swaps forward to it.
- omm runStart a chat with an installed model - Ollama chats right here in the terminal, KoboldCpp/text-generation-webui start with the model loaded, GUI apps are opened.
- omm scanSummarize memory, model storage, and installed local AI runners.
- omm searchSearch curated models, cached candidates, and HuggingFace by name.
- omm setting11 sub-commandsView or change omm settings (telemetry, upload policy, version, calibration, catalog trust).
- omm setupRe-run the first-time setup wizard (hardware scan + engine checklist).
- omm tuneRecommend context, GPU offload, threads, and batch size for a model.
- omm uninstallalias: rmUninstall a model and clean up all symlinks/manifests. Pass `all` to uninstall every model installed via omm.
- omm unlinkRemove one or more models' links from one runner (or every runner with --runner all) without touching the hub file or links into other runners.
- omm unpinUndo `omm pin`: a future `omm install --force` stops archiving this model, and any version already archived for it is deleted. Run `omm rollback` first if you might still want that archived version.
- omm updateReinstall omm from the latest source and refresh its data.
- omm upgradealias: upLook for a better model than the ones already installed - a newer curated successor, or a higher-quality quantization from the same repo that still fits this machine's memory - and install the ones you approve. With no argument (or `all`), scans every model installed via omm.
- omm verifyProve that an installed model can load and return local text.
Exported from omm 0.3.101.