Run NameQuick Fully Local
Rename files with a local AI model via Ollama or LM Studio on Mac and Windows, or MLX on Apple Silicon Macs. No API key, no cloud AI — file analysis stays on your computer.
Self-Managed works with your own API key (BYOK) or with a local model. This page covers the local option: NameQuick names files with a model on your own computer, on Mac or Windows. Use a local model served by Ollama or LM Studio, or MLX on an Apple Silicon Mac, and file contents are analyzed on your own hardware, with nothing sent to a cloud AI provider. No API key and no per-rename cost.
This is the right setup if you rename confidential material (contracts, medical or financial documents, client files) or simply want a one-time purchase with zero running costs.
Quality depends on your computer and the model you choose
Local models are smaller than cloud models and run on your own hardware, so results can vary. They are slower than cloud models and small ones produce simpler filenames. On Apple Silicon with a mid-size model, quality is very usable for documents; for image-heavy work you'll need a vision-capable model. Try your real files during the trial and judge for yourself.
Which engine to use#
| Engine | Mac | Windows | Setup |
|---|---|---|---|
| Ollama (recommended start) | Yes | Yes | The easiest way to run local models. Install it once. |
| LM Studio | Yes | Yes | A graphical app for browsing and running models. Install it and turn on its local server. |
| MLX | Apple Silicon, macOS 14 or later | No | Apple's machine-learning framework, optimized for M-series chips. Built in: NameQuick manages the runtime, with no separate server app to install. |
MLX is not available on Intel Macs or Windows; use Ollama or LM Studio there.
Set up the engine#
Open SettingsAI and select Ollama, LM Studio, or MLX. There is no server address to enter.
Ollama isn't installed
Click Download Ollama ↗, install it, and come back to NameQuick. For LM Studio, install the app, then select LM Studio in Settings → AI.
“Ollama isn't running”
Start it from NameQuick. Turn on Start the engine automatically if you want NameQuick to start it whenever a rename needs it.
“LM Studio's server is off”
Open LM Studio, go to the Developer tab, and turn the local server on.
“isn't answering — the port may be blocked”
Allow network access when Windows asks and click Restart engine.
Pick a model#
Pick a model under Recommended and click Get, or choose one you already pulled under Your models. NameQuick marks how well each model fits your computer.
| Label in the app | What it means |
|---|---|
| Runs well on your device | Fits your computer well. |
| Tight fit on your device | A tight fit for your computer. |
| Text only — scans need a model that reads images | Handles text documents only. |
| Reads scans & photos | Pick this for documents and images. |
If your model is text-only, NameQuick can still rename documents via extracted text, but images and scanned PDFs need a vision model. You'll see a message telling you so.
Run a first test#
Select engine and model
Choose the engine and model in SettingsAI.
Rename a test file
Rename a test file to confirm everything works.
Local vs. cloud at a glance#
| Local (Ollama / LM Studio / MLX) | Cloud (your own key) | Managed | |
|---|---|---|---|
| Privacy | Analysis stays on your computer | Sent to your chosen provider | EU cloud, not used for AI training |
| Cost per rename | Free | Fractions of a cent, billed by provider | Renames from $12/mo |
| Setup | Install a model runner once (MLX is built in) | Create a key once | None |
| Speed & quality | Depends on your computer and model | Highest | Highest |
| Works offline | Yes, once set up (license check every 7 days) | No | No |
Caution
Naming works offline once the engine and model are set up. Installing engines, downloading models, and license and update checks still use the internet, and the license needs a successful online check at least every 7 days.
You can switch providers anytime in SettingsAI. For exactly what leaves your computer in each mode, see Privacy & Your Data.
Common issues#
NameQuick can't reach the model
Settings → AI shows what is missing. Start Ollama or LM Studio from NameQuick, turn on LM Studio's local server in its Developer tab, or, if you see isn't answering — the port may be blocked, allow network access when Windows asks and click Restart engine.
Images or scans aren't being renamed
Your model is text-only. Pull a vision-capable model and select it.
Renames are slow
Bigger models are slower; try a smaller one, or close other heavy apps. Check the fit label, for example Runs well on or Tight fit on your device. Apple Silicon is dramatically faster than Intel here.
More in Troubleshooting.