Description
Run high-performance AI models like Qwen locally on your iPhone and iPad. Offline. Private. No login. No data collection. Powered by Apple MLX.
Meet Drissy, the on-device personal assistant that doesn't just chat—it browses. While other apps stop at text generation, Drissy uses the fine-tuned Drissy-Qwen3.5-2B model to read webpages and provide grounded, factual answers while keeping your identity 100% anonymous.
Why Drissy?
Total Privacy by Design: No accounts, no tracking, and no cloud syncing. Your search queries and AI data never leave your device.
New Voice Mode: Get answers faster than ever. Speak naturally to Drissy to start research, ask questions, or summarize data without touching the keyboard.
Choose Your Brain: Now supporting all major on-device models. From the efficient Drissy-Qwen3.5-2B to the cutting-edge 1-bit quantized Bonsai 8B, you can pick the model that perfectly balances speed and power for your device.
Smart "Browse for You" Feature: Unlike standard local LLMs, Drissy can search the web, read the top results, and synthesize information into a single clear answer—all locally.
Optimized for Apple Silicon: Specifically built using Apple MLX to ensure lightning-fast responses on your iPhone and iPad.
100% Offline Intelligence: Access core AI capabilities and your entire research history even without an internet connection.
Threaded Research Context: Drissy remembers the flow of your investigation. Ask follow-up questions and dive deeper into complex topics without ever losing the thread.
Pure Performance: An uncompromising, ad-free experience with no bloated interface. Just pure, private intelligence at your fingertips.
Search Smarter, Not Louder
Read the Web: Stop clicking through dozens of tabs. Let Drissy read the articles for you and provide a concise summary.
Anonymity Guaranteed: The ultimate alternative to data-hungry, cloud-based search engines and AI assistants.
Download Drissy now and experience the next generation of private, on-device AI.
Nouveautés (v1.8.4)
1. Added openrouter llm api fallback for those devices which aren't technically feasible yet to run llms
2. Added stats to every on-device ai response, to shows tokens per sec, time taken, tokens etc