LocalRAG! icon
Ask your documents. Check the source.

Import PDFs, Office docs, EPUBs, images, audio, and video. Original files, extraction, indexing, and search stay on your device. Use Claude for higher-quality answers. With Local AI selected, answer generation stays on your device and can work offline once prepared.

Download on the App Store GET IT ON Google Play

Available on iPhone, iPad, and Android

A 1-week free trial may be available when shown as eligible in the App Store or Google Play.

LocalRAG! chat answer with source citations

What's New

v3.7

Complete collection answers, steadier Local AI, and Claude Sonnet 5.5

Collection-wide questions can now process every document with visible progress and coverage. Local AI adapts context and memory use to the selected model and device. New on-device models are LFM2.5 2.6B, LFM2.5 8B (MoE), and Gemma 4 E4B on iOS, plus LFM2.5 1.2B on Android. Model downloads resume in the background, and Android can continue answer processing outside the foreground. Answer-completeness checks, multilingual retrieval, and Markdown rendering are also improved.

Collection-wide Q&ALocal AI reliabilityBackground processingClaude Sonnet 5.5Multilingual retrieval

Your knowledge, searched locally. Your choice of AI.

LocalRAG! separates document handling from answer generation, so you can see exactly what stays on your device.

01

Your knowledge

PDF, Office, EPUB, images, audio, and video — 25 formats in one collection.

02 · ON DEVICE

Local RAG

Extraction, indexing, and retrieval run on your device. The original file is not uploaded.

03 · Choose the answer engine
Local AI

Once the model is ready, document search, questions, and answer generation work on-device without an internet connection.

OR
Cloud AI

Choose Claude for stronger answers. Only your question and retrieved excerpts are sent — not the original file.

04

A cited answer

Open the cited page or timestamp and verify the answer against the source.

Start with the work you need to finish

The same local knowledge pipeline becomes a different solution depending on the documents you add and the question you ask.

02

Review contracts and policies

Locate clauses, obligations, and source passages.

See every use case →

On-Device AI vs. Claude API

Two ways to chat with your documents. Choose what fits your needs.

On-Device AIClaude API
Internet requiredNo — fully offlineYes
Document & question data sent to serverNoneQuestion + relevant excerpts only
Where answers are generatedOn your deviceClaude API
Model downloadRequired before local answersNot required
PlatformiOS & AndroidiOS & Android

Both modes process documents (text extraction, indexing, search) entirely on your device. The on-device AI takes this further by generating answers locally too — so your documents, your questions, and the answers all stay on your device. Separately from either mode, the app sends usage statistics — tied to a random per-install ID, never to you, and never including your content; our Privacy Policy lists exactly what those contain.

See the product, not a promise

Ask. Read the answer. Open the source.

LocalRAG! keeps the retrieval path visible, so an answer is a starting point for verification rather than a black box.

Open the full demo page

The essentials, without the feature maze

25 formats, cited answers, collections, OCR, transcription, and your choice of Local or Cloud AI.

📄

25 Formats Supported

Documents (PDF, EPUB, DOCX, XLSX, PPTX, TXT, MD, CSV, RTF, HTML), images (JPG, PNG, HEIC, WebP), audio (MP3, M4A, WAV, AAC, FLAC) and video (MP4, MOV, M4V) — plus scanned and copy-protected PDFs.

💬

Natural Language Chat

Ask questions about your documents in plain language. Get answers with source citations.

📍

Inline Citations & PDF Jump

AI answers show [1][2] source references you can tap. Jump directly to the cited page with yellow highlight on the exact passage.

🔒

Privacy First

Original files, extraction, indexing, and search stay on your device. With Claude selected, the app sends your question and retrieved excerpts, not the original file.

🔍

BM25 + Cross-Language Search

Okapi BM25 search with on-device embedding model (E5 Multilingual) for semantic search across languages. Ask in Japanese, find answers in English documents.

🖼

Image OCR + Editor

Import photos of whiteboards, receipts, or scanned documents. On-device OCR extracts text — and you can edit the extracted text before indexing.

📴

Offline AI Mode

With Local AI selected, document content, questions, and answers stay on-device and can work offline after the local model is available.

🤖

Model Selection

Use Local AI for an on-device workflow, or choose one of the Claude models available in the app.

📂

Collections (Folders)

Organize documents in nested folders up to 3 levels. Parent scope includes all child documents for focused Q&A.

PDFEPUBDOCXXLSXPPTXTXTMDCSVRTFHTMLJPG/PNGHEIC/WebPScanned PDFMP3/M4AWAV/FLACMP4/MOV

Frequently Asked Questions

Is LocalRAG! truly local? What data is sent to the cloud?

Document processing — text extraction, chunking, and search indexing — happens on your device, and the original file is not uploaded. With Local AI selected and a local model prepared, the question and answer generation also stay on-device. With Claude selected, the app sends the question and retrieved excerpts to the API, not the original file. Separately, the app sends anonymous usage statistics that do not include file contents, question text, or answer text.

How is this different from ChatPDF, PDF.ai, or other cloud-based tools?

LocalRAG! keeps extraction, indexing, and retrieval on your device, so the original file is not uploaded. You can prepare Local AI for on-device answers, including offline use, or choose Claude, which receives the question and retrieved excerpts. In either mode, answers include source references so you can return to the imported material and verify them.

Does Anthropic use my data for AI training?

Anthropic states that inputs and outputs from its commercial API are not used to train generative models by default. Standard API retention may be up to 30 days, subject to Anthropic's current terms and any applicable agreement. LocalRAG! sends only the question and retrieved excerpts in Claude mode, not the original file.

Can I use LocalRAG! without a subscription?

Yes. The free plan includes 5 questions per day with Claude Haiku and the on-device AI. You can also bring your own Anthropic API key (BYOK) for unlimited access to all models at no subscription cost — you only pay Anthropic directly for API usage.

How does the offline AI (local LLM) work?

LocalRAG! can prepare a language model on your iPhone, iPad, or Android device. Model size and performance vary by the model and device. Once a compatible model is ready and Local AI is selected, document search, questions, and answer generation can work without an internet connection. Anonymous usage statistics are handled separately and do not include your content.

What document formats are supported?

LocalRAG! supports 25 formats across four categories: 11 document formats (PDF, TXT, MD, CSV, RTF, HTML, HTM, EPUB, DOCX, XLSX, PPTX), 6 image formats (JPG, JPEG, PNG, HEIC, HEIF, WebP) processed with on-device OCR, 5 audio formats (MP3, M4A, WAV, AAC, FLAC), and 3 video formats (MP4, MOV, M4V) transcribed on-device with WhisperKit (iOS) and whisper.cpp (Android). Scanned and copy-protected PDFs are also supported.

Start chatting with your documents

Download on the App Store GET IT ON Google Play

Start free with 5 questions/day, or subscribe for more.
Free, Standard ($4.99/mo), and Pro ($9.99/mo) plans available. Cancel anytime.
Terms of Use · Privacy Policy