Loading...
Run Meta: Llama 4 Maverick in an invisible overlay that listens to your meetings, sees your screen, and answers before you need it — no meta-llama account, no API keys.
The actual overlay running Meta: Llama 4 Maverick — live scenarios, real UI. Always on top of every app, invisible on screen shares.
The same overlay, two ways to use it — and Meta: Llama 4 Maverick is one tap away in both.
Meta: Llama 4 Maverick reads what it sees — screenshots, charts, photos, whiteboards, PDFs. Capture the screen, drag-select a region, or turn on Use image so every message carries a fresh screenshot for it to analyze.
Pluely transcribes your mic and system audio in real time while Meta: Llama 4 Maverick also sees your shared screen — so the suggested answer accounts for the slide, the chart, or the doc everyone is looking at.
Download for your platform, browse release history, or explore our development journey
Apple Silicon & Intel
x64 Architecture
Debian Package
Loading releases...
Latest release downloads
Browse all releases
Development timeline
Download Pluely now and experience the privacy-first AI assistant that works seamlessly in the background.
Open the overlay's model picker, search for Meta: Llama 4 Maverick, and your next question runs on it. No meta-llama account, no keys to manage — it's bundled with your Pluely plan, and you can switch models mid-conversation.
The overlay never shows up in screen shares, recordings, or screenshots, and no bot joins your calls. Your chats, meetings, and files stay in a local, encrypted database on your device.
No meta-llama account, no API key, no setup ritual.
Click the model picker in the bottom-right corner of the overlay's composer.
Tap it in the list — it's bundled with your plan, so there's nothing to connect or configure.
Your next answer runs on Meta: Llama 4 Maverick, and you can switch again mid-conversation with context intact.
Meta: Llama 4 Maverick is one of the models bundled with Pluely Pro: available in the overlay's picker with zero setup, and switchable mid-conversation. Here's how meta-llama describes it:
Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward...
Install Pluely and it's one tap away — for the next error, the next call, the next thing you'd rather not figure out alone.
Free plan forever · Invisible on screen shares · No bot joins your calls