Your private AI.
Right on your phone.
Scan documents, ask questions, get answers — all powered by AI that runs entirely on your device. No cloud. No accounts. Works offline.
Free to start · Pro from $9.99/mo · Founding pricing for waitlist members
See it in action
Your documents.
Your device. Only.
Process sensitive documents, get sourced answers, and extract structured data — entirely on-device. Nothing ever touches a server.
UnQuestSummarize the key clauses in my NDA
Based on your NDA document, here are the key clauses:
Confidentiality Period: 3 years from disclosure date
Non-Compete: 12-month restriction within 50mi radius
IP Assignment: All work product belongs to Company
Privacy First
Cloud AI sees everything.
We see nothing.
Your data stored on their servers
Your data stays on your device. Always.
API calls for every query
Zero network calls. Verifiably.
Prompts logged and analyzed
Zero telemetry on your content
Data used to train their models
Your data trains nothing but your own knowledge graph
Trust their privacy policy
Trust the architecture: on-device, encrypted, auditable
How it works
Three steps. Zero cloud.
Download once. Runs offline.
Install UnQuest and a compact 2.8GB AI model. After that, zero internet needed. No accounts, no API keys, no cloud — it all runs on your phone.
Feed it your world
Drop PDFs, scan contracts with your camera, import notes. UnQuest chunks, embeds, and indexes everything locally into a private knowledge graph.
Put it to work
Analyze an NDA. Categorize a month of receipts. Track trends across lab reports. Sourced answers with page attribution — all in airplane mode.
Founding member pricing for waitlist · No credit card required
Capabilities
Everything runs on your device
No API keys. No accounts. No data leaving your phone. Intelligence that processes, understands, and reasons — right where your data lives.
Document Intelligence
Analyze a 30-page contract in under a minute. Extract clauses, summarize terms, flag risks — entirely on your phone, even in airplane mode.
Knowledge That Compounds
Every document you process makes UnQuest smarter. A persistent knowledge base and cross-conversation memory connect insights across all your files.
Camera to Knowledge
Photograph a receipt, scan a contract, snap a lab report. Instant OCR, indexing, and AI analysis — one tap from paper to searchable knowledge.
Sourced Answers
Ask anything about your documents. Get answers with exact page and paragraph attribution — so you can verify every claim, not just trust it.
On-Device Voice
Speak to your knowledge base. On-device speech-to-text (Parakeet) and natural text-to-speech — your voice is never streamed to a server.
Local Image Generation (soon)
On-device image generation with the Bonsai MLX model is in active development. When it lands, prompts and pixels stay on your phone — no cloud render.
Siri & Shortcuts
Ask UnQuest straight from Siri, and automate document workflows with the Shortcuts app via native App Intents.
Domain Workflows
Purpose-built intelligence for professionals: NDA clause extraction, expense categorization, medical trend tracking. Real work, not generic chat.
Zero-Knowledge Architecture
Zero network calls for inference, zero telemetry, zero third-party SDKs, AES-256 encryption at rest. Keys in the Secure Enclave. A guarantee, not just a policy.
Models
Run the best open models — locally.
Download Qwen, Gemma, Llama, DeepSeek, Phi and more, optimized for Apple Silicon. Complex work routes to your downloaded model; quick tasks go to Apple Intelligence — automatically.
Qwen 3.5
1.2B–9B · vision · default
Gemma 3 / 3n
1B–4B · multimodal
Llama 3.2
1B–3B · 128K context
DeepSeek R1
1.5B · reasoning
SmolLM3
3B · dual reasoning
Phi-4 Mini
3.8B · math & logic
IBM Granite
2B · tool-use
LiquidAI LFM2.5
1.2B · fastest
Bonsai (MLX)
image gen · soon
Apple Intelligence
iOS 26+ · no download
Qwen 3.5 4B
DefaultState-of-the-art reasoning in 2.8GB. Handles contract analysis, document Q&A, and structured extraction at 15-25 tokens/sec.
Parameters
4B
Quantization
Q4_K_M
Size
2.8 GB
Context
262K native
Apple Intelligence
iOS 26+Apple's built-in foundation model. No download, instant response. Handles classification, simple chat, and triage — freeing the big model for heavy lifting.
Download
None
Tasks
Chat, classify
Speed
Instant
Requires
iOS 26+
Smart Inference Routing
Simple chat & classification
→ Apple Intelligence
RAG, analysis & workflows
→ Your downloaded model
Document processing & OCR
→ llama.cpp / MLX
How we compare
Privacy and capability.
Not either/or.
Cloud AI has the features but leaks your data. Other local AI apps protect your data but only offer chat. UnQuest gives you both.
Cloud AI = ChatGPT, Claude, Gemini · Other local apps = PocketPal, Enclave, Private LLM, Locally AI
Compatibility
Built for Apple Silicon
Native Swift. Metal GPU acceleration. Purpose-built for on-device inference — not a web wrapper, not a port.
iOS
17.0+
RAM
8 GB minimum
Storage
~3 GB for model
Chip
A17 Pro / A18+
FAQ
Questions answered
Completely. Once you download the 2.8GB model, every computation runs on your device's GPU. Analyze contracts, process receipts, query your knowledge base — all in airplane mode. Zero network calls, zero telemetry, zero third-party SDKs. Verifiably private.
ChatGPT and Claude send your data to remote servers for processing. UnQuest runs entirely on your device — your prompts, documents, and responses never leave your phone. Beyond privacy, UnQuest also processes your documents, builds a knowledge graph, and gives sourced answers with page attribution. The trade-off: the on-device model is smaller (4B vs 100B+), so it excels at focused tasks like document Q&A, extraction, and analysis.
Most on-device AI apps are chat interfaces that run a model locally. UnQuest goes further: it ingests and indexes your documents, builds a persistent knowledge graph that connects insights across files, and runs purpose-built workflows for legal, financial, and medical use cases — all with the same on-device privacy guarantee.
Drop in a PDF or scan with your camera. UnQuest extracts text (OCR for scanned docs), chunks it intelligently, generates embeddings, and indexes everything into a private, searchable knowledge graph — on-device. Then ask questions and get sourced answers with exact page attribution. Run workflows: extract NDA clauses, categorize receipt line items, track trends across lab reports.
All stored data — documents, conversations, embeddings, knowledge graph — is encrypted with AES-256-GCM. The encryption key is stored in the Secure Enclave via Apple's Keychain. Even if someone extracts your phone's storage, the data is unreadable without biometric authentication.
Any iPhone with 8GB RAM and Apple Silicon — that's iPhone 15 Pro, all iPhone 16 models, and the entire iPhone 17 lineup. Standard iPhone 15 and older models don't have enough memory to run a 4B parameter model. iPad and Mac support is planned.
15-25 tokens per second on A18 Pro with Metal GPU acceleration. Responses start streaming immediately — you don't wait for the full answer. Apple Intelligence handles quick tasks (classification, triage) near-instantly, freeing the heavy model for deep analysis.
Voice is fully on-device today: the Speech tab does speech-to-text (Parakeet, with Apple Speech as a fallback) and natural text-to-speech, so you can talk to your knowledge base hands-free. On-device image generation (Bonsai MLX) is in active development and ships in a later update. Either way, nothing leaves your phone.
Yes. UnQuest ships native App Intents, so you can ask it from Siri and build automations in the Shortcuts app — all running against on-device models.
There's a free tier with on-device chat, voice, and a few document imports. Pro is $9.99/month or $79.99/year and unlocks unlimited documents, RAG with citations, the full model catalog, and image generation. Professional workflow packs (Legal, Financial, Medical) are one-time purchases — $9.99 each, or $19.99 for all three.
We will never route your content through external servers. That's an architectural guarantee, not just a policy. Future plans include encrypted iCloud sync (using your own private CloudKit database) for multi-device access, but inference and document processing will always remain on-device.
Pricing
Start free. Go Pro when you're ready.
A simple Pro subscription unlocks the full workspace. Workflow packs are one-time add-ons. No cloud costs — because there's no cloud.
Free during beta · Pricing takes effect at launch · Waitlist members get founding pricing
Free
Get started with on-device AI chat, voice, and basic document processing.
- Unlimited AI chat (on-device)
- On-device voice mode (speech-to-text + TTS)
- 3 document imports
- Basic summarization
- 10 saved conversations
- Works 100% offline
Pro
or $79.99/year — save 33%
The full private AI workspace. Cancel anytime — your data always stays on-device.
- Everything in Free
- Unlimited documents
- Ask questions about your docs (RAG + citations)
- Knowledge base + semantic search across files
- Cross-conversation memory
- Full on-device model catalog
- Siri & Shortcuts
- AES-256 encryption at rest
Workflow Packs
Purpose-built intelligence for professionals. One-time purchase, add the packs you need.
- Legal: NDA analysis, clause extraction
- Financial: receipts, expense reports, tax docs
- Medical: lab reports, trend tracking
- Structured data extraction
- Domain-specific prompts
Your AI should be
yours alone
Private on-device intelligence is coming soon. Join the waitlist for early access and launch pricing.
Free to start · Pro from $9.99/mo · Waitlist members get founding pricing
No spam. We'll email you once when it's ready.