An open API service indexing awesome lists of open source software.

https://github.com/muhittincamdali/swiftintelligence

Privacy-first modular AI toolkit for Apple developers with Vision, NaturalLanguage, Speech, benchmarks, and release proof.
https://github.com/muhittincamdali/swiftintelligence

ai-toolkit apple apple-developer-tools benchmarking coreml ios macos natural-language on-device-ml privacy privacy-first speech swift swift-ai swift-package-manager tvos vision-framework visionos watchos

Last synced: 2 months ago
JSON representation

Privacy-first modular AI toolkit for Apple developers with Vision, NaturalLanguage, Speech, benchmarks, and release proof.

Awesome Lists containing this project

README

          


SwiftIntelligence hero banner


SwiftIntelligence is an Apple-native AI workflow stack for teams that need
Vision, NaturalLanguage, Speech, privacy controls, benchmarks, and release proof to ship as one product path.


Use it when multiple Apple AI capabilities must ship together with proof discipline.
Skip it when raw Apple APIs or a narrower package already close the product path cleanly.


Start Here
Examples Hub
Release Proof


Need rival-fit first? See Competitive Matrix.
Need the human trust router? See Trust Start.
Need the repo story in another language? See README Languages.

## First Decision

| Need | Best next move |
| --- | --- |
| one maintained Apple workflow that spans modules and release proof | [Five-Minute Success Path](Documentation/Getting-Started.md#five-minute-success-path) |
| exact trust, readiness, and release-claim boundaries | [Trust Start](Documentation/Trust-Start.md) |
| named-rival fit before adoption | [Competitive Matrix](Documentation/Comparisons/Competitive-Matrix.md) |
| only one low-level framework call or a specialist package | stay on raw Apple APIs or choose the narrower package first |

Use [Public Proof Status](Documentation/Generated/Public-Proof-Status.md) only when exact claim wording matters.

## Demo Gallery



SwiftIntelligence IntelligentCamera flagship demo screenshot

`IntelligentCamera` is the fastest honest proof of value in this repository.
It closes the strongest maintained path: `Vision -> NLP -> Privacy`.

Flagship media:
- [Guide](Examples/DemoApps/IntelligentCamera/README.md)
- [Recording](Documentation/Assets/Flagship-Demo/intelligent-camera-run.mp4)
- [Caption](Documentation/Assets/Flagship-Demo/caption.txt)
- [Media Policy](Documentation/Assets/Flagship-Demo/README.md)

```bash
bash Scripts/validate-flagship-demo.sh
```




SmartTranslator secondary demo screenshot

SmartTranslator
NLP -> Privacy -> Speech


Best secondary path for text analysis, tokenized preview, and translated output.



Guide
Recording
Media Policy





VoiceAssistant secondary demo screenshot

VoiceAssistant
NLP -> Privacy -> Speech


Best secondary path for assistant-style intent handling and privacy-aware response output.



Guide
Recording
Media Policy



`SmartTranslator` and `VoiceAssistant` carry real media and maintained guides.
They are not flagship proof surfaces, but they are no longer source-only demos.

Use [Examples Hub](Examples/README.md) when you need the canonical demo chooser.

## Product Surface


SwiftIntelligence capability board

| Product lane | Start with demo | Compare first | Why this lane matters |
| --- | --- | --- | --- |
| Vision | [IntelligentCamera](Examples/DemoApps/IntelligentCamera/README.md) | [Vision vs Apple Vision](Documentation/Comparisons/Vision-vs-AppleVision.md) | strongest maintained product path and flagship evaluator route |
| NaturalLanguage | [SmartTranslator](Examples/DemoApps/SmartTranslator/README.md) | [NLP vs NaturalLanguage](Documentation/Comparisons/NLP-vs-NaturalLanguage.md) | clearest text-first path for summaries, entities, and transformed output |
| Speech | [VoiceAssistant](Examples/DemoApps/VoiceAssistant/README.md) | [Speech vs Apple Speech](Documentation/Comparisons/Speech-vs-AppleSpeech.md) | assistant-style response flows without pretending to be a speech-recognition leader |
| Privacy | [IntelligentCamera](Examples/DemoApps/IntelligentCamera/README.md) | [Privacy vs CryptoKit + Security](Documentation/Comparisons/Privacy-vs-CryptoKit-Security.md) | makes tokenization and safer handling visible in app-facing workflows |
| Benchmarks | [Examples Hub](Examples/README.md) | [Benchmark Baselines](Documentation/Benchmark-Baselines.md) | keeps claims inside thresholds, manifests, and immutable bundles |
| Release Proof | [Trust Start](Documentation/Trust-Start.md) | [Latest Release Proof](Documentation/Generated/Latest-Release-Proof.md) | closes trust and release questions with generated evidence |

## Competitive Reality

SwiftIntelligence does **not** win by pretending to be `MLX`, `coremltools`, or a generic cross-platform inference runtime.

It wins only if it is clearly better at:

- Apple-native workflow composition
- proof-backed release discipline
- multi-module adoption guidance
- faster first success for real app teams

Current top comparison set:

- [SwiftIntelligence vs Top Rivals](Documentation/Comparisons/Competitive-Matrix.md)
- [NLP vs Apple NaturalLanguage](Documentation/Comparisons/NLP-vs-NaturalLanguage.md)
- [Vision vs Apple Vision](Documentation/Comparisons/Vision-vs-AppleVision.md)
- [Speech vs Apple Speech](Documentation/Comparisons/Speech-vs-AppleSpeech.md)
- [Privacy vs CryptoKit + Security](Documentation/Comparisons/Privacy-vs-CryptoKit-Security.md)

## Architecture Signal


SwiftIntelligence architecture board

| Layer | Maintained surface |
| --- | --- |
| Flagship workflows | `IntelligentCamera`, `SmartTranslator`, `VoiceAssistant` |
| Perception | `SwiftIntelligenceVision`, `SwiftIntelligenceML` |
| Language | `SwiftIntelligenceNLP`, `SwiftIntelligenceSpeech`, `SwiftIntelligenceReasoning` |
| Trust and transport | `SwiftIntelligencePrivacy`, `SwiftIntelligenceNetwork`, `SwiftIntelligenceMetrics`, `SwiftIntelligenceCache` |
| Release operating system | `SwiftIntelligenceBenchmarks`, generated proof pages, release bundles, media pack |

## Trust Surface


SwiftIntelligence trust and distribution board

| Signal | Current truth |
| --- | --- |
| Publish readiness | `ready` |
| Distribution posture | `release-grade` |
| Required release device floor | `Mac + iPhone` |
| Flagship media | `published` |
| Secondary demo media | `SmartTranslator, VoiceAssistant published` |
| Canonical trust start | [Trust Start](Documentation/Trust-Start.md) |
| Latest immutable proof | [Latest Release Proof](Documentation/Generated/Latest-Release-Proof.md) |

## Docs Map

| You are... | Start here |
| --- | --- |
| Evaluating the repo | [Documentation Hub](Documentation/README.md) |
| Trying the strongest path | [Getting Started](Documentation/Getting-Started.md) |
| Comparing options | [Competitive Matrix](Documentation/Comparisons/Competitive-Matrix.md) |
| Reviewing trust and release proof | [Trust Start](Documentation/Trust-Start.md) |
| Looking for product examples | [Examples](Examples/README.md) |
| Contributing or maintaining | [Contributing](CONTRIBUTING.md) |

## Install The Strongest Path

```swift
dependencies: [
.package(url: "https://github.com/muhittincamdali/SwiftIntelligence.git", from: "1.2.2")
],
targets: [
.target(
name: "MyApp",
dependencies: [
.product(name: "SwiftIntelligenceCore", package: "SwiftIntelligence"),
.product(name: "SwiftIntelligenceVision", package: "SwiftIntelligence"),
.product(name: "SwiftIntelligenceNLP", package: "SwiftIntelligence"),
.product(name: "SwiftIntelligencePrivacy", package: "SwiftIntelligence")
]
)
]
```

Validate the same public path the repo claims:

```bash
swift build
bash Scripts/validate-flagship-demo.sh
bash Scripts/validate-examples.sh
swift test
```

## Module Surface

Core modules and what they do

| Module | Role |
| --- | --- |
| `SwiftIntelligenceCore` | shared configuration, logging, runtime utilities |
| `SwiftIntelligenceML` | on-device training, prediction, evaluation, cache management |
| `SwiftIntelligenceNLP` | sentiment, entities, keywords, summaries, topics |
| `SwiftIntelligenceVision` | classification, detection, OCR, segmentation, enhancement |
| `SwiftIntelligenceSpeech` | speech-related types, synthesis, voice catalogs |
| `SwiftIntelligencePrivacy` | tokenization, anonymization, secure storage, compliance helpers |
| `SwiftIntelligenceReasoning` | higher-level reasoning primitives |
| `SwiftIntelligenceNetwork` | network-layer helpers |
| `SwiftIntelligenceCache` | cache primitives |
| `SwiftIntelligenceMetrics` | metrics and observability support |
| `SwiftIntelligenceBenchmarks` | benchmark runners, artifacts, thresholds, baselines |

## Community and Maintenance Surface

| If you need to verify... | Start here | Why |
| --- | --- | --- |
| how contributions are judged | [Contributing](CONTRIBUTING.md) | the quality bar for code, proof, examples, and public claims |
| how vulnerabilities should be reported | [Security Policy](SECURITY.md) | private intake path and current security operations matrix |
| how to get help or route an issue | [Support](SUPPORT.md) | support channels, repro expectations, and maintainer gate set |
| how GitHub-facing metadata stays aligned | [GitHub Distribution](Documentation/GitHub-Distribution.md) | About box, topics, README story, and public proof routing |
| how releases are prepared and validated | [Release Process](Documentation/Release-Process.md) | release asset, proof, and publication discipline |
| how to support ongoing maintenance | [GitHub Sponsors](https://github.com/sponsors/muhittincamdali) | direct maintenance support linked from `.github/FUNDING.yml` |
| how collaboration stays safe | [Code of Conduct](CODE_OF_CONDUCT.md) | conduct expectations across public interactions |

## License

Released under the [MIT License](LICENSE).