Run large language models entirely on your Mac. No cloud, no API keys, no telemetry. Download, open, chat.
macOS 13+ · Apple Silicon required
Features
Fllint ships everything in a single folder. No scattered configs, no background daemons, no hidden caches.
Everything runs locally. No data ever leaves your machine — no cloud calls, no telemetry, no tracking.
Models, conversations, config — all in one folder. Delete it and Fllint is gone. Zero system footprint.
Download a model from within the app and start chatting. No terminal, no configuration, no guesswork.
Send images to vision-capable models. Attach PDFs with optional OCR text extraction.
Connect to Ollama or any OpenAI-compatible server for access to even more models.
Full control over inference parameters, context size, GPU layers, and system prompts when you need it.
How it works
No terminal. No setup wizard. No 20-step guide.
Grab the latest release from GitHub. Unzip and drop the Fllint folder wherever you want — Desktop, Applications, an external SSD.
Double-click Fllint.app. It starts a local server and opens in your browser. A menu bar icon shows it's running.
Pick a model tier — Lite for quick answers, Standard for balanced quality, Pro for maximum capability. The download starts in-app.
Requirements
Bigger models are smarter but need more memory. Start with Lite and upgrade when you're ready.
| Model Tier | Size | RAM | Best For |
|---|---|---|---|
| Lite | ~2 GB | 8 GB+ | Quick answers, lighter hardware |
| Standard | ~9 GB | 16 GB+ | Balanced quality and speed |
| Pro | ~22 GB | 32 GB+ | Maximum capability |
Open source. Free forever. Your data stays yours.
Download for macOS