KernelAI vs LocalRAG!

ON-DEVICE AI Downloaded compatible models run locally on your device. KernelAI checks model compatibility and memory needs before you download, so you can choose a model that fits your iPhone or iPad. CHAT WITH DOCUMENTS Ask questions about PDFs, DOCX, CSV, Markdown, and text files. Summarize, search, and find useful passages with on-device document chat. UNDERSTAND IMAGES Attach a photo, screenshot, chart, or diagram and ask what it shows with a compatible vision model. Image analysis runs on-device. 50+ MODEL OPTIONS Choose from 50+ models, including Llama, Gemma, Qwen, DeepSeek, Mistral, Phi, and more. Import compatible GGUF models from Files, URLs, or your local network. WRITE, LEARN, AND CODE Use local models for writing, brainstorming, explanations, and code. Generate, explain, and review code with models selected for programming. CUSTOMIZE YOUR AI Set a system prompt and generation parameters for each model. See tokens-per-second, context usage, and generated thinking output from supported reasoning models. VOICE, SIRI, AND SHORTCUTS Dictate prompts with voice input. Ask KernelAI from Siri or build it into Shortcuts on supported iOS versions. On supported iOS 26 devices, Apple's built-in on-device model can answer in the background; downloaded models may open KernelAI while they run. ORGANIZE YOUR WORK Keep conversations in folders, search your history, branch a chat, edit or regenerate messages, and export your work. Use KernelAI on iPhone and iPad, including from the Share Sheet and Home Screen widget. OPTIONAL WEB SEARCH Turn on web search when you need current information. Add your own Tavily or Exa API key; web search is off by default. Provider fees may apply. TEN APP LANGUAGES Use KernelAI in English, French, German, Spanish, Russian, Brazilian Portuguese, Japanese, Korean, Traditional Chinese or Arabic. Choose your language during onboarding or change it in Settings. Arabic includes a right-to-left interface. FREE TO USE KernelAI remains free. Local AI works offline after downloading a compatible model. No subscription is required for this version. Privacy Policy: https://kernelai.app/privacy Terms & Conditions: https://kernelai.app/terms

LocalRAG! — Any file. One conversation. PDF, Excel, Word, PowerPoint, EPUB, images — and now audio recordings and video files. No matter the format, just add them to a collection. AI searches across all your documents at once. No cloud uploads required. NEW IN v3.0: AUDIO & VIDEO TRANSCRIPTION Import MP3, M4A, WAV, AAC, FLAC audio files and MP4, MOV, M4V video files. WhisperKit transcribes everything entirely on your device — no audio ever leaves your phone. AI-powered cleanup automatically removes filler words and stutters for cleaner transcripts. Tap any answer to jump straight to the timestamped segment. HOW PEOPLE USE IT • Meeting recordings turned into searchable knowledge Import your meeting recordings. Ask "What was decided about the budget?" Get answers with timestamps you can jump to. • Lecture videos and study materials together Mix lecture videos, PDFs, and your notes in one collection. Ask "Explain the method in Chapter 3" — AI searches text and transcripts. • Interview recordings stay private Journalists, researchers, lawyers — your sensitive recordings are transcribed entirely on-device. Nothing is uploaded. • All your project docs in one place PDFs, Word minutes, Excel spreadsheets, audio briefings — formats don't matter. Just ask "What was decided?" • Search across languages English PDFs, Japanese spreadsheets, Thai notes — language doesn't matter. Ask in English, get answers from Japanese documents. 25 FORMATS SUPPORTED Documents (PDF, EPUB, DOCX, XLSX, PPTX, TXT, MD, CSV, RTF, HTML), images (JPG, PNG, HEIC, WebP) with on-device OCR, audio (MP3, M4A, WAV, AAC, FLAC) and video (MP4, MOV, M4V) with on-device WhisperKit transcription. Scanned and copy-protected PDFs also supported. PRIVACY FIRST Text extraction, transcription, indexing — all happen on your device. Your original files and audio recordings never leave your device. With the local LLM option, your questions and answers stay on the device too — no document text, no question, no answer is ever transmitted. With Claude API mode, only your question and relevant text snippets are sent. No account, no email, no advertising IDs, no third-party trackers. The app does send usage statistics so we can improve it — they never identify you and never include your content, and our Privacy Policy lists exactly what they contain. ANSWERS WITH SOURCES AI automatically finds relevant sections and responds with source citations — document name and page number, or audio file with timestamp. Tap a citation to jump straight to it. Summarize, translate, analyze — any instruction works. COLLECTIONS Organize documents into folders. Switch chat scope by collection for focused Q&A sessions. Supports hierarchical structure (up to 3 levels). BUILT-IN VIEWERS View PDFs, Excel, PowerPoint, EPUB, images, and audio/video transcripts directly in the app with thumbnail previews and timestamp navigation. MODEL SELECTION Use your own API key to choose between Claude Sonnet 5, Haiku 4.5, or Opus 5. Or switch to the built-in local LLM to keep your documents and questions entirely on-device. Balance speed, accuracy, and cost to suit your needs. 8 LANGUAGES Available in English, Japanese, German, Spanish, French, Portuguese, Korean, and Chinese. SUBSCRIPTION - 1-week free trial included - Full access to all features during trial - Subscription begins automatically after trial ends - Cancel anytime SUBSCRIPTION DETAILS - Monthly and yearly plans available - Payment is charged to your Apple ID account at confirmation - Subscription auto-renews unless cancelled at least 24 hours before the end of the current period - Manage or cancel anytime in Account Settings Terms of Use: https://localrag.app/terms Privacy Policy: https://localrag.app/privacy