Cortiq: Local AI Models

Cortiq turns your phone into a fully local AI workstation. Chat with large language models that run entirely on your device — no cloud, no account, no data collection. Your conversations never leave your phone.

WHY CORTIQ

• 100% private — models run on-device, works completely offline
• Free and open source, no subscriptions, no tracking
• Powered by CMF (Cortiq Model Format), an efficient format for mobile AI

CHAT

• Streaming replies with per-message stats: latency, tokens, tokens/sec
• Persistent chat topics — every conversation keeps its own context
• Attach documents (txt, md, code, json and more) to your prompt
• Markdown rendering, copy, regenerate, stop
• "Disable thinking" mode — reasoning models answer directly

MODEL LIBRARY

• Local model library with full metadata: architecture, quantization, context size
• RAM fit check before loading — a warning instead of a crash
• One-tap load and unload to free memory and battery
• Import .cmf files from device storage

HUGGING FACE CONVERTER

• Search the Hugging Face Hub, pick a model and quantization, watch live progress
• On-device multi-threaded conversion of safetensors models to CMF
• Quantizations from full F16 down to ternary (~2-bit) for the smallest footprint
• Featured mobile-ready models download in one tap
• Resilient downloads: automatic retry and resume after network drops

LOCAL AI SERVER

• Serve loaded models to your network over an OpenAI-compatible API
• Use your phone as an AI backend for laptops and other devices
• QR code for instant connection setup

Cortiq is part of the open CMF ecosystem. Model performance depends on your device: newer phones with more RAM run larger models faster.