OpenIntelligence · Version 5.3 · Requires iOS 26 · Lifetime on sale until September 29
Ask your files anything. Inspect every answer.
Turn PDFs, Office documents, code, images, audio, and video into private, searchable libraries. Reading, indexing, retrieval, ranking and verification all run on your device, so answers work with no connection at all. And when your files don't support an answer, the app says so and stops instead of writing something plausible.
- iPhone · iPad · Mac
- No account or API key
- Data Not Collected
- Citation-backed answers
Built for the questions that matter
What can you do with it?
Find exact answers
Ask a direct question and jump back to the supporting passage, page, timestamp, or file.
Understand complex material
Turn dense reports, manuals, transcripts, and research into a grounded explanation.
Compare multiple sources
Build one evidence-backed answer across an entire private library instead of one file at a time.
Work beyond PDFs
Search Office files, code, images, audio, and video alongside ordinary documents.
Verify the result
Inspect citations, route receipts, and confidence cues, or an abstention when evidence is weak.
Your documents stay on your device, and every answer shows where it came from.
Ask questions across your own PDFs, Office files, notes, scans, code and recordings. Retrieval and answering run on your iPhone, and every answer carries the passages it came from so you can check it.
OpenIntelligence ingests PDFs, Office files, code, images, audio, and video into private libraries, then plans, retrieves, reranks, verifies, and cites the evidence used for each answer. Ingestion, indexing, retrieval, and verification all stay on-device, and the answer itself is written by the Apple Intelligence models already built into the device. Nothing is sent to a third-party AI provider. On iOS and macOS 27, longer evidence-heavy questions can optionally use Apple Private Cloud Compute; the app shows exactly what would be sent and asks you to approve it first.
Capabilities
What Makes OpenIntelligence Different
Local-first by architecture
Parsing, indexing, retrieval, reranking, and evidence verification happen on-device. The app remains useful offline through its on-device answer path.
More than PDFs
PDFKit, Vision OCR, structured parsers, and media extraction paths handle digital documents, scanned pages, code, images, audio, and video.
Retrieval that can explain itself
SQLite FTS5, local vector indexing, query analysis, reranking, routing, and context packing are orchestrated as a real agentic retrieval loop instead of a single-vector-only search path.
Answers you can inspect
The app is built to show evidence, drop unsupported claims, and refuse when the documents do not support a confident answer.
Private library boundaries
Private libraries, retrieval controls, and diagnostics make it possible to scope work and inspect how the answer path behaved.
Execution policy control
Every answer identifies its route, read from an execution receipt rather than from what was requested. On iOS and macOS 27, Private Cloud Compute is limited to final synthesis after disclosure and consent, with retrieval and verification staying local either way. On older systems every route is local.
Under the hood
Technical Profile
Core APIs
- Platforms: iPhone, iPad, and Mac
- Parsing: PDFKit, Vision OCR, structured document processing
- Answer paths: on-device synthesis using the Apple Intelligence models already on the device, or consent-gated Private Cloud Compute on iOS and macOS 27
- Retrieval: agentic RAG with SQLite FTS5, local vector indexing, reranking, and evidence packing
Data boundary
- Always local: Libraries, normalized text, indexes, retrieval, reranking, verification, and citations
- Leaves only with consent: Final synthesis through Apple Private Cloud Compute on iOS and macOS 27, after payload disclosure and approval
- Not the product: generic chatbot behavior over arbitrary cloud context
- Receipt: Every answer surfaces the route used
Workflow
How OpenIntelligence Runs
Import into private libraries
Files are assigned to local libraries and normalized through text extraction or OCR, depending on document quality.
Build lexical and vector indexes
The app preserves structure where possible, then stores lexical signals and vector representations for hybrid retrieval.
Plan, retrieve, and verify evidence
Query planning, reranking, routing, and evidence packing are used to keep the answer path tied to source material rather than generic model guesswork.
Use Apple Intelligence to answer or abstain
The app answers through the selected execution route, attaches citations and confidence cues, or abstains when the evidence is not strong enough.
Go deeper
Go Deeper
Interactive technical showcase
See the full retrieval trace, execution routes, supported file types, changelog, and product FAQ at fascinaiting.me.
Support URL
https://gunzino.me/openintelligence/support
Privacy Policy URL
https://gunzino.me/openintelligence/privacy
Your files. Your evidence. Your choice of route.