How much is OpenIntelligence?

OpenIntelligence is listed for sale at $3,000 USD.

What are OpenIntelligence's current metrics?

Monthly visitors: Undisclosed. Monthly revenue: Undisclosed.

OpenIntelligence - Sell Project

Listed May 3, 2026

About OpenIntelligence

Private Apple Intelligence Foundation Models RAG engine built for Apple Silicon

OpenIntelligence OpenIntelligence is a private Apple-native document-intelligence engine that currently lives inside a shipped iOS codebase. The public app is the proving ground. The thing for sale is the engine itself: the Swift/iOS code, the current source-distributed SDK lane, the evaluation materials, and the handoff context around it. This started because I kept wanting AI help with the exact documents I did not want pushed through a generic hosted workflow. Manuals. IFUs. Service guides. Internal references. Training material. Product docs. Sensitive operational stuff. So I built the Apple-native version I wanted to exist. What is actually here A repo-root Swift package surface for OpenIntelligenceEngine A small public engine facade under OpenIntelligence/SDK/OpenIntelligenceEngine.swift A repo-side consumer sample app under Samples/SourceSDKHost/ Validation scripts for package build, sample build, and simulator smoke tests Buyer and partner packet zips rebuilt from current source A packet-local evaluation host app Benchmark harness code, reporting scripts, and pipeline documentation What the engine does It is not a thin Foundation Models wrapper. The engine handles document import, text extraction, OCR fallback, structure-aware chunking, metadata enrichment, local SQLite full-text indexing, local vector indexing, hybrid retrieval, reranking, nearby-context expansion, context packing, grounded answer generation, citations, and verification-oriented response behavior. Apple Foundation Models is the generation layer inside that stack, not the whole stack. What I am actually selling The strongest commercial lane today is a private source-distributed SDK with assisted integration, packet-based diligence, and a clean handoff path. There is no hosted web API in the buyer packet. Documents and indexes are local-first by default. I am not representing this as a polished self-serve enterprise SDK, a toolchain-agnostic binary SDK, or a regulated-use answer system. This is also not a sale of the whole app business by default. It is a sale or license discussion around the engine, the repo, the evaluation materials, and the handoff context around the engine. If a repo snapshot is transferred, some app-side files still exist around the engine because that is where the engine was built, but those are context around the asset, not the asset being valued. Current state The engine is real and substantial The source-distributed SDK lane is real and validated The buyer packet and partner packet are current The repo is a serious head start for Apple-native private document QA It still needs more cleanup and productization before it looks like a polished standalone SDK

Project details

Built with: Swift,JavaScript,CSS,Shell,Ruby,Python,HTML

Website: https://github.com/Gunnarguy/OpenIntelligence

Video 1: https://www.loom.com/share/9860865ae137493bb87ebf7fa44ce28e

Next step: If I were buying this from someone else, I would think about it as a head start, not a finish line. First thing I’d suggest is looking into Apple’s LoRA - seems remarkable and I don’t have the hardware to make it possible. The most obvious next move is engine separation. The valuable logic already exists, but parts of it are still coupled to app services, runtime paths, storage assumptions, diagnostics, and UI surfaces. A buyer could turn that into a cleaner SDK or internal framework with a narrower API and a more reproducible handoff path. After that, I think the biggest gains are in retrieval and source fidelity. The engine already has ingestion, OCR-oriented processing, chunking, local full-text search, vector retrieval, context packing, Apple Foundation Models integration, source review, verification-oriented behavior, and benchmark tooling. What still needs work is getting the exact right supporting chunk or page more reliably, preserving important evidence through packing, and getting the system better at refusing when support is weak. The other big area is hard documents. Manuals, IFUs, service guides, spec-heavy PDFs, tables, and procedural material are exactly why I built this, and they are also the hardest things to get right. That means table extraction, page-level evidence, procedural-step fidelity, cross-reference handling, and exact-value lookup are all good places for a buyer to keep pushing. I also think the benchmark side matters more than the demo side. The benchmark harness gives a buyer a way to test the engine on their own corpus instead of just watching me run a polished sample. For the right buyer, that is where the real value starts showing up. It would need fine-tuning for whichever domain the buyer intends to aim it towards. The strongest buyer path is probably not trying to make this work for every document on earth. It is to narrow it around one real workflow where grounded answers over private docs are actually worth money. <ol><li data-list="bullet">Field-service manuals</li><li data-list="bullet">Internal technical references</li><li data-list="bullet">compliance material</li><li data-list="bullet">Product support libraries</li><li data-list="bullet">Training workflows</li><li data-list="bullet">Private knowledge bases on Apple devices</li></ol> That is the opportunity as I see it: take the real ingestion, indexing, retrieval, context prep, answer generation, source review, and benchmark work that already exists here, then harden it around one actual buyer problem until it becomes much more than a prototype.

For sale

Asking price: $3,000 USD

What's included

Source code, All related design work

Included in the sale or license discussion by default: Private engine/source-code repo snapshot or agreed repo transfer scope pinned to a commit Repo-root Swift package surface for OpenIntelligenceEngine Current public engine facade under OpenIntelligence/SDK/OpenIntelligenceEngine.swift Engine-relevant Swift source across document processing, chunking, embeddings, storage, vector search, retrieval, verification, and answer orchestration layers Repo-side source consumer sample app and simulator smoke-test path Buyer packet zip Partner packet zip Packet-local evaluation host app Benchmark harness code, reporting scripts, and evaluation fixtures already in the repo Limited handoff documentation Limited post-transfer clarification if written into scope Incidental repo context only, not part of the asset story: Some app-side project files remain in the same repo because the engine currently lives inside that tree. Default handoff path: Scope and exclusions are written down. The buyer packet and partner packet are used for diligence. The transfer is pinned to a named commit SHA. Delivery happens via repo transfer or repo archive plus both packet zips. A short handoff call is held. The buyer confirms receipt of access and artifacts. Not included by default: App Store app transfer App Store listing or release pipeline consumer app business context Apple Developer account access signing certificates provisioning profiles API keys or service credentials personal accounts private benchmark corpora sensitive benchmark traces long-term support custom feature development exclusivity production SDK warranty regulated-use validation Separately negotiable: App Store app transfer exclusive field-of-use license broader source-code acquisition longer transition support custom integration help private walkthrough sessions deeper SDK cleanup and packaging work

Reason for selling

I feel that the project requires skills that I do not necessarily have and would like to see someone else take over.

Project metrics

Average Monthly Visitors: Undisclosed

Average Monthly Revenue: Undisclosed

Average Monthly Downloads: Undisclosed