Ogma AI: Offline AI Chat

Ogma AI: Your 100% Offline, Private AI Assistant Experience true AI privacy. Ogma AI runs open-source artificial intelligence models directly on your iPhone and iPad — 100% offline, zero cloud processing, zero tracking, and no account required. Your conversations, documents, and photos remain entirely on your device. Whether you are traveling on an airplane, commuting without cellular reception, or working with confidential data, Ogma AI provides intelligent, on-demand assistance anywhere, anytime. KEY FEATURES ON-DEVICE OPEN-SOURCE LLMS Download and chat with leading open-weight models directly from the built-in catalog: • Compatible with top open model architectures including Llama, Gemma, DeepSeek, Qwen, Mistral, and Phi. • Powered by an optimized llama.cpp inference engine built natively for Apple Silicon. • Advanced Flash Attention and KV-cache quantization for fast generation and minimal memory consumption. • Import and run your own custom GGUF model files directly from the Files app. OFFLINE DOCUMENT & PDF CHAT (RAG) Analyze documents securely on your device: • Import PDFs, Microsoft Word (DOCX), Excel spreadsheets (XLSX, CSV), and plain text files. • 100% local vector embeddings and retrieval (sqlite-vec) with zero network traffic. • Summarize reports, extract key insights, and query complex documents with cited references. ON-DEVICE VISION INTELLIGENCE Analyze images without uploading them to external servers: • Run vision-language models on-device to understand photos, diagrams, and screenshots. • Extract text from receipts, transcribe documents, or analyze layout structures locally. HANDS-FREE REAL-TIME VOICE MODE Interact with your AI naturally using voice: • On-device speech recognition for fluid, continuous input. • Natural text-to-speech audio playback. • Dedicated full-screen voice interface designed for hands-free productivity. INTERACTIVE ARTIFACTS Transform generated code into live visual components: • Render interactive HTML, SVG diagrams, and scripts directly in your conversation. • Inspect UI prototypes, visualize charts, or preview code output in an isolated sandbox. CUSTOM ENDPOINTS & CLOUD INTEGRATION (OPTIONAL) Connect your own services when needed: • Connect directly to cloud providers (OpenAI, Anthropic Claude, OpenRouter) using your own API key. • Link to your self-hosted Ollama or vLLM server over local Wi-Fi. • Core offline AI functionality is always available; cloud provider connections, full-screen hands-free voice mode, and premium themes are Supporter features. PRIVACY BY DESIGN • No account, registration, or sign-in required. • Zero telemetry: No third-party analytics SDKs, advertising IDs, or tracking. • Local storage: Conversations and index vectors are stored strictly in local SQLite storage. • Rich presentation: Full support for Markdown, LaTeX math equations, and syntax-highlighted code. • Personalization: 21 color themes (including OLED Pure Black), 18 fonts, and 12 background patterns. • Multilingual: Interface and chat support for 39 languages. LEGAL & PRIVACY • Terms of Use (EULA): https://weareapexcreators.github.io/policies/ogma_t_c.html • Privacy Policy: https://weareapexcreators.github.io/policies/ogma_pp.html • Apple Standard EULA: https://www.apple.com/legal/internet-services/itunes/dev/stdeula/