OGAM
Run GGUF LLMs, vision, speech, and stable diffusion completely offline on mobile or Mac.
Offline GGUF text, vision, voice, and image generation with GPU and NPU acceleration.
OGAM is an offline AI suite that runs GGUF language models, Stable Diffusion, and Whisper speech-to-text natively on your mobile device or Mac. It suits professionals who need localized text generation, document analysis, and voice tools without exposing data to external servers.
What it is
The app handles on-device text generation using architectures like Llama and Qwen, processes PDFs via local vector embeddings, and executes Stable Diffusion image generation. It manages RAM allocation explicitly and supports local-network servers and tool calling.
Who it's for
It suits mobile-first professionals, translators, and writers who require strict offline operation and hardware acceleration. It does not suit those seeking cloud-hosted scalability or web-based multi-user management.
Key features
Hardware Acceleration
Utilizes phone GPUs and NPUs via OpenCL and Metal for faster token rates.
Local Knowledge Base
Embeds uploaded PDFs and text documents locally in SQLite using a bundled MiniLM model.
On-Device Voice and Vision
Integrates Whisper speech-to-text and multimodal models like SmolVLM directly on hardware.
Explicit RAM Management
Provides policy controls to monitor and override memory limits for resident models.
Up and running in 3 steps
Our honest verdict
What's great
- Runs entirely offline with zero data leaving your device
- Supports hardware acceleration across Snapdragon GPUs/NPUs and Apple Silicon
- Includes local PDF chunking and vector retrieval via SQLite
- Connects to local-network OpenAI-compatible servers like Ollama
Where it falls short
- Requires powerful mobile hardware for acceptable token generation rates
- Pro tier requires a paid lifetime license or monthly subscription for advanced features
- Experimental NPU acceleration may cause model failures or fallback to CPU
Pricing & license
The core app is free under its open-source license. An optional Pro tier adds voice mode, custom personas, tool integrations, and sync for $119 lifetime or $7.99 monthly.
Open remote AI-training roles
767 open remote AI-training roles, updated November 14, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Does any data leave my device during use?
No. All text generation, speech transcription, vision analysis, and image creation run locally on your hardware.
What hardware is required for optimal performance?
Flagship devices with Snapdragon 8 Gen 2/3 or Apple Silicon provide the best token rates and NPU image generation speeds.
